Problem Solving and Process Improvement

Tools and Techniques Guide Book

Max Zornada, Henley Management Group Pty Ltd, A.C.N. 054 337 101 61 Carrington Street, Adelaide 5000 South Australia, Tel (08) 8237 0586, Fax (08) 8237 0555, Email:[email protected] Web: www.hmg.com.au

Problem Solving and Process Improvement Tools and Techniques

Page 2

Contents
1. Introduction ......................................................................................................................................... 5 2. Problem Solving and Decision Making in Organisations ................................................................... 6
Conventional Approaches to Problem Solving .................................................................................................. 6 Problem Solving and Improvement in Six Sigma Organisations ....................................................................... 8 Root Cause Analysis ........................................................................................................................................ 10 The Problem Solving Tools and Techniques .................................................................................................... 13

3. Process Mapping ............................................................................................................................... 15

What is Process Mapping? .............................................................................................................................. 15 When to use Process Mapping ......................................................................................................................... 16 How to Build a Process Map ........................................................................................................................... 17 Issues to consider when developing Process Maps .......................................................................................... 18 Variations on Basic Process Mapping ............................................................................................................. 19

4. Problem Identification Tools and Techniques .................................................................................. 22

Brainstorming .................................................................................................................................................. 23 List Reduction .................................................................................................................................................. 27 Nominal Group Technique............................................................................................................................... 28 Multivoting ....................................................................................................................................................... 31

5. The Seven Tools of Quality .............................................................................................................. 32

Cause and Effect Analysis (Fishbone Diagram) .............................................................................................. 33 Check Sheets .................................................................................................................................................... 38 Pareto Analysis ................................................................................................................................................ 40 Histograms ....................................................................................................................................................... 46 The Scatter Diagram ....................................................................................................................................... 51 Run Charts ....................................................................................................................................................... 57 Control Charts ................................................................................................................................................. 60 The X Chart...................................................................................................................................................... 65 The X-bar and R Chart .................................................................................................................................... 68 The i and mr Chart........................................................................................................................................... 73 The np Chart. ................................................................................................................................................... 79 The p Chart ...................................................................................................................................................... 84 The c Chart ...................................................................................................................................................... 88

Problem Solving and Process Improvement Tools and Techniques

Page 3

The u Chart ...................................................................................................................................................... 92 Interpreting Control Charts ............................................................................................................................. 97 Process Capability ......................................................................................................................................... 100

6. Extensions to the Seven Tools of Quality....................................................................................... 102

Interviews ....................................................................................................................................................... 103 Questionnaires ............................................................................................................................................... 106 Dot Plots ........................................................................................................................................................ 109 Box Plots ........................................................................................................................................................ 111 Force Field Analysis ...................................................................................................................................... 121

7. The Seven Quality Management and Planning Tools .................................................................... 124

The Affinity Diagram ..................................................................................................................................... 126 Relations Diagram ......................................................................................................................................... 132 The Systematic Diagram (Tree Diagram) ...................................................................................................... 138 The Matrix Diagram ...................................................................................................................................... 142 The Decision Matrix ...................................................................................................................................... 145 The Prioritisation Matrix ............................................................................................................................... 149 The Allocation Matrix .................................................................................................................................... 152 The Process Decision Program Chart ........................................................................................................... 155 The Network Diagram (Critical Path) ........................................................................................................... 158

Appendix: Understanding Data and Variation ................................................................................... 159

1. Introduction to Variation ........................................................................................................................... 160 2. Observing Variation .................................................................................................................................. 160 3. Characterising Data .................................................................................................................................. 160 4. Understanding Variation ........................................................................................................................... 170

Problem Solving and Process Improvement Tools and Techniques

Page 4

1. Introduction
This Tools and Techniques guidebook is intended to act as a reference source for problem solving and process improvement practitioners, facilitators, team leaders and team members. In this guidebook, you will find a comprehensive coverage of all the major tools and techniques that have been found to be useful in problem solving and process improvement applications. This includes the original seven tools proposed by Dr. Kaoru Ishikawa as well as many others used in support of these. Before looking at each tool or technique in detail, the early part of this handbook addresses the context in which the tools are used, by considering the issue of problem solving and decision making in organisations, and contrasting conventional approaches to structured quality based approaches. In particular, we will look at the DMAIC problem solving process and process improvement framework that is part of the Six Sigma philosophy.

Problem Solving and Process Improvement Tools and Techniques

Page 5

2. Problem Solving and Decision Making in Organisations

Conventional Approaches to Problem Solving
Being regularly confronted with problems to be solved, issues to be resolved and decisions to be made is and has always been a major part of the day-to-day work routine for all employees, regardless of the type and size of organisation, or the level at which they work. The only apparent difference as one moves through the organisational levels from shopfloor to boardroom, is the type and scope of problems, issues, decisions and the individuals power to address them. Individual experts and expertise, troubleshooters and gut feel, usually in the management and executive levels have dominated problem solving and decision making approaches in most organisations. The result of such approaches have been knee-jerk reactions to symptoms, temporary fixes that fail to address the true causes of a problem and lack of commitment by the those affected to implement the solution. Hierarchical organisational structures and lengthy internal lines of communication means that only a small proportion of all problems and issues are ever known about by the management levels who have the authority to address them. Therefore, not only is problem solving and decision making overly dependent on a few key individuals, but even where the individuals are effective problem solvers, they rarely find out about more than a relatively small proportion of the problems. The consequences of such approaches are that most of the organisations employees work in an environment where they have to put up with a whole host of unresolved issues and problems, which they know about but are helpless to act upon, as part of their day-to-day work life. This leads to frustration and demotivation. A more dangerous consequence is that when particular unresolved issues and problems persist over a long period of time, they gradually become accepted as part of the organisations business as usual paradigm. Both management and workforce become blind to the problem or issues, accepting them as a natural part of the way things are.

Problem Solving and Process Improvement Tools and Techniques

Page 6

A common response to any outsider or newcomer challenging or highlighting such a problem or issue is often - thats the way we do things around here!. In this manner, problems, inefficiencies and bad practices become institutionalised. Attempts to address them are met with resistance, often coming from those who would benefit most from the change. Such organisations are easy prey for quality driven competitors. When faced with the need to improve, they are usually blind to the opportunities within their organisation and so focus their energies blaming external factors.

Problem Solving and Process Improvement Tools and Techniques

Page 7

Problem Solving and Improvement in Six Sigma Organisations

Problem solving in Six Sigma organisations follows a structured process that provides the discipline to ensure problem solvers do not jump to solutions before they have analysed the causes and that the solutions when implemented, really do address and eliminate the root causes of problems. In particular, there is a strong focus on reducing variation and the cost of poor quality. The core of the Six Sigma approach is the Define, Measure, Analyse, Improve, Control cycle or DMAIC. This is illustrated in Figure 2.1 below.

Control

Define

Improve

6 Measure

Analyse

Figure 2.1. The DMAIC Cycle

When applying the DMAIC cycle, problem solving and process improvement teams work through each stage in turn, where by they are required to: Define the problem/opportunity, customer and customer requirements and the process that will be affected; Measure Determine and implement the required process measured to monitor and establish the current performance baseline for the process and the problem impact. Analyse Analyse the problem and the process by conducting a root cause analysis until the potential root cause(s) is found.

Problem Solving and Process Improvement Tools and Techniques

Page 8

Improve Generate potential solutions and assess them suitability, before selecting the preferred solution and implementing it. Control institutionalise the change by embedding it into the process through procedures, systems and standards. Implementing appropriate performance measures for ongoing monitoring of the process.

The basic DMAIC process can be further expanded to provide detailed guidance for problem solving and process improvement teams by presenting the key stages with their detailed steps in the form of a story board the team work their way through, as shown in Figure 2.2.

Figure 2.2: DMAIC Problem Solving Process Improvement Storyboard

Problem Solving and Process Improvement Tools and Techniques

Page 9

Root Cause Analysis

One of shortfalls of conventional problem solving has often been that implementation of the recommended solution either does not solve the problem or only temporarily solves the problem. ie. the problem reoccurs some time later.

Causes and Root Causes

This is usually because the proposed solution may have been one that addressed the symptoms only and did not resolve the true or root cause of the problem. What we observe when we identify a problem are usually one or more symptoms attributable to some cause. When we problem solve, we attempt to identify the cause so that we may eliminate it. Unfortunately, the cause may itself be due to some other cause. Resolving this initial cause is a reaction to the symptom and will not resolve the problem. The observed symptom or problem may temporarily (or permanently) disappear only to remanifest itself later in the same or a different shape or form. However, if we continue to work our way through the chain of causes and effects, we shall eventually find the root cause or the true cause of the observed symptom. Resolving the root cause not only resolves the problem, but also eliminates the remainder of the non-root causes in the chain. This relationship is illustrated in Figure 2.3.
O b s e r v e d P r o b l e m S y m p t o m C a u s e

TheProblemSolvingProces
Figure 2.3. Causes and Root Causes
Problem Solving and Process Improvement Tools and Techniques

C a u s e

R o o t C a u s e

Page 10

Quality based problem solving focuses on identifying and eliminating root causes. It does this by providing a structured process through the DMAIC cycle, whereby causes can be tested to determine whether or not they are root causes, as they are found. Root causes can be further assessed to determine whether they are the correct root cause for the problem at hand. ie. we may find a root cause, but it may not be the root cause of our problem.

Identifying Root Causes

An overview of the root cause analysis process can be presented as follows.

O b s e r v e d s y m p t o m " T h e P r o b l e m "

A p p l y p r o b l e m s o l v i n g p r o c e s s C a u s e f o u n d ?

N o t a r o o t c a u s e

N O

I s c a u s e a c t i o n a b l e ?

Y E S C a u s e f o u n d ?

N o t t h e r i g h t c a u s e

N OD o e s a c t i o n ( s )
f i x p r o b l e m ?

Y E S

P r o b l e m S o l v e d
Figure 2.4 Finding the Root Cause A key feature of non-root causes, is that they are rarely directly actionable. Problem solvers may often derive an action by jumping to conclusions and acting on these.

Problem Solving and Process Improvement Tools and Techniques

Page 11

If cycling through the root cause analysis process identifies a cause that is not directly actionable, it is unlikely to be a root cause. This cause then becomes the focus of the next cycle of analysis. The process is repeated until an actionable cause is found. An actionable cause will be a root cause, but it may be the wrong root cause. Therefore, the process requires problem solving teams to check whether or not actioning the root cause actually solves the problem. If it doesnt, the process is repeated until the right root cause is found.

Using the tools and techniques

Although there are many tools and techniques available, in practice, one finds that a few tend to be used time and time again. The common cycle of events, linked to the specific tools used most of the time is illustrated in Figure 2.5.

Define the problem

Cause & Effect Analysis (Fishbone diagram) used to brainstorm possible causes. Vote on most likely and develop data collection approach and instrument. This will usually be some sort of checksheet. Collect data using the checksheet or other appropriate data gathering proforma. Analyse Data using a Pareto chart, determine most likely cause. If cause not actionable, redefine problem to determine causes of this cause and repeat the process.

Is cause actionable?

No

Yes
Figure 2.5 Using the Problem Solving Tools

Problem Solving and Process Improvement Tools and Techniques

Page 12

The Problem Solving Tools and Techniques

There are now a wide range of tools and techniques available to the quality practitioner. The problem solving and continuous improvement processes previously described all call for the use of specific tools at the various stages of the process. A key technique is the use of flow charting or process mapping as a tool for understanding business and work processes. By far the most common tools and techniques used to supplement process mapping are the original seven tools of quality put forward by Dr. Kaoru Ishikawa. These are: Check sheets; Cause-and-Effect diagrams; Pareto Charts; Histograms; Scatter Diagrams; Run Charts; Control Charts. Since then, many more tools and techniques have been added to the growing arsenal that is now available to the quality practitioner. Most of the tools and techniques have been developed for use at both the individual level or as part of a group or team process. In addition, their use has been extended well beyond their initial manufacturing application as the quality philosophy has spread to the broader business community. This booklet introduces the reader to: Process Mapping using the flowcharting technique; Tools used to support initial problem identification. This includes: Brain Storming Nominal Group Technique Multivoting

Problem Solving and Process Improvement Tools and Techniques

Page 13

 The original seven tools of Quality. Namely: Histograms; Cause-and-Effect diagrams; Check sheets; Pareto Charts; Scatter Diagrams; Run Charts; Control Charts. Tools that have been added and accepted as supplements to the seven tools. Including: Interviews; Opinion Questionnaires; Dot Plots; Box Plots; Force Field Analysis; Process Capability Study; The seven quality management and planning tools developed to support and guide managerial problem solving and decision making. These include: Affinity Diagram; Interrelationship Diagram; Systematic Diagram; Matrix Diagram; Matrix-Data Analysis; Process Decision Program Chart; Activity Network Diagram.

Problem Solving and Process Improvement Tools and Techniques

Page 14

3. Process Mapping
What is Process Mapping?
Process mapping is a technique used to analyse business and work processes. Process mapping or flow charting as it is sometimes called, allows users to build a picture or map of a process - a process map, using easy-to-understand visual symbols. A process map reveals what a process actually looks like, its shape and the flow of work through it. The symbols most often used in process mapping/flow charting are given in Figure 3.1.
Symbol Name Operation or process step Brief Definition Depending on the level of detail being developed, can be used to denote anything from a simple task, major activity or a whole sub-processes. Point where a decision must be made before any further action can be taken. Point at which a form or report is generated by the process. Used to indicate the process is continued elsewhere on the flow diagram or on another sheet. Optionally used to describe an input or output from a processing block. Use to connect all blocks to display the sequence in which operations are performed. Used to indicate the start and end of a process.

Decision Point

Document Generated Continuation Point Input/Output Block Flow lines Termination point

Figure 3.1 Process Mapping Symbols Figure 3.2 illustrates part of a typical process map. The process mapping technique, when properly applied, is particularly good at illustrating the structure and logic behind the process flow and highlighting key decision points.

Problem Solving and Process Improvement Tools and Techniques

Page 15

S t a r t
a i s e o f f i c i a l o r d e r P r o c e s s s t e pR
S e n d t o D a t a P r o c e s s i n g

D o c u m e n t g e n e r a t e d

Y e s D e c i s i o n p o i n t I s O r d e r O K ? N o

C a l l A c c o u n t M a n a g e r
Y e s I s A c c o u n t M g r i n ? N o

P r o v i d e A d d i t i o n a l I n f o

S e n d t o A c c o u n t M g r .

P r o v i d e A d d i t i o n a l I n f o R e t u r n t o D a t a P r o c

C o n t i n u a t i o n p o i n t

2
Figure 3.2. A Process Map - this example shows the order entry part of an order fulfilment process

When to use Process Mapping

Process mapping is used whenever there is a need to understand the process. This is usually the first step in problem solving and process improvement situations. Process maps will readily highlight: inherent structural inefficiencies built into a process; areas offering the greatest potential for improvement; what sensible measures of performance are likely to be. Process mapping is the key tool to used gain an understanding of business or work processes as a prerequisite to any process improvement or process reengineering activities.

Problem Solving and Process Improvement Tools and Techniques

Page 16

How to Build a Process Map

1. Select the process. Determine which process you wish to study. 2. Define the boundaries. Define the boundaries to your process. Will you be looking at a small work process that is part of a much larger business process or will you be looking at a major process that connects with external suppliers at one end and external customers at the other? Define where you process begins and where it finishes. 3. Identify the start and end points. Identify where the process starts and where it ends, identify the inputs to and outputs from the process. In particular, identify what it is that causes the process to operate or triggers the start of the process and what is it that determines when the process has concluded. ie. what event, occurrence or condition. 4. Start drawing the process map. To do this: Identify all the process steps that must occur between the beginning and end points of the process. ie. identify each activity, task, decision point, inspection, check point etc. that occurs; Lay them out in the sequence in which they occur in practice. Use the standard symbols to depict the different types of activities that may occur; Connect the various steps with flow lines to show how the process flows through the steps. 5. Check accuracy and completeness. Ensure all the team members involved in developing the process map agree on what the process looks like. If not already represented on the team, ensure people involved with each aspect of the process are given the opportunity to comment. In particular, check that every step and decision point has been included and that the chart depicts what actually happens and not what team members think should happen.(a common trap). 6. Prepare the final process map 7. Repeat step 5. If necessary, make final adjustments.

Problem Solving and Process Improvement Tools and Techniques

Page 17

Issues to consider when developing Process Maps Level of Detail

The level of detail required for a process map will be dependent on the use the team wish to make of the map. Is a simple overview all that is required or do you need a detailed flowchart to identify areas of waste and inefficiency. Typically, when looking for opportunities to change fundamental process structures, an overview which illustrates the existing structure may be all that is required. Where a team is scrutinising a process to: identify opportunities to reduce waste, complexity and inefficiency; identify key control points; develop meaningful performance measures;

a much more higher level of detail will need to be developed.

Who should be involved?

Ideally, the process map should be constructed by a team, whose membership includes individuals from each part of the organisation involved with the process. When process mapping is left to management, staff experts or outside consultants, the end result rarely provides a true reflection of what actually happens. Rather, it is often the individuals perceptions of the process or the way it should operate. It is essential to get the input of the people that have to work in the process day in-day out to get the true picture of the way the process actually works.

Problem Solving and Process Improvement Tools and Techniques

Page 18

Variations on Basic Process Mapping Process Blocking

In some situations it may not be necessary to document a process in detail. All that may be required is a basic understanding of the process structure. In the case of large business wide processes, management may need to get an understanding of the overall process shape, before deciding to allocate teams to examine the various subsections in detail - ie. develop detailed process maps. Process blocking is a tool that can be used to do this. In process blocking, we represent major chunks of the process as blocks without detailed flowcharting of what goes on inside the block. An example of a companys order fulfilment process, depicted using the process blocking technique is shown in Figure 3.3.
S t a r t

R e c e i v e O r d e r

P u r c h a s i n g P r o c e s s

R a i s e O f f i c i a l O r d e r

D a t a P r o c e s s i n g

C r e d i t C h e c k

W a r e h o u s e p i c k i n g p r o c e s s

D i s p a t c h & D e l i v e r y

R e j e c t i o n p r o c e s s

A c c o u n t P a y a b l e

I n s t a l l & C o m m i s s i o n

F i n i s h

Figure 3.3. Order Fulfilment Process - Process Blocks

Problem Solving and Process Improvement Tools and Techniques

Page 19

Organisational/Functional Process Maps

An example of this variation of a process map is illustrated in Figure 3.4. The Organisational/Functional Process Map, which I refer to as a Zig-Zag chart because of the way process flows often zig-zag across the chart, requires the team to identify all the functional positions in the organisation which are involved in a process. All the people involved in a process, commencing with the customer at the far left, all internal positions and any external suppliers, usually placed at the far right are placed horizontally across the top of the chart. The process map is then drawn so as to depict which parts of the process are performed by whom.
F r o n t C u s t o m e r L i n e o c u m e n t ' n B r o k e r a l u a t i o nA s s e s s m ' tD S t a m p S e t t l e m ' t M a n a g e r V

Figure 3.4. A Zig-Zag Chart of a Bank Lending Process

Problem Solving and Process Improvement Tools and Techniques

Page 20

Depending on the size of the process and the level of detail required, this type of chart may be developed for the detail process map, for process or for some combination of the two which gives an appropriate level of detail. The advantages of this technique are that it: Highlights how the people in the organisation work together, or are required to work together to make the process work; Highlights where activities happen, are required to happen or may be organised to be done in parallel;

Readily highlights inefficiency and opportunities for inefficiencies to arise which manifests itself in the form of zig-zagging and handovers of between positions. The other major advantage of this approach, in the cases of service processes which interface with external customers, is that it allows the team to understand how the customer sees the process. The interaction between the customer and the organisation occurs every time a process flow line crosses the boundary between the customer column and the organisation. Regardless of what the process does inside the organisation, these will be the parts of the process the customer sees and will form the basis of their view of how the organisation delivers its service.

Problem Solving and Process Improvement Tools and Techniques

Page 21

4. Problem Identification Tools and Techniques

In many cases, the problem to be addressed is obvious. At other times, the problem may not be all that apparent or well defined. The problem identification tools and techniques are all aimed at improving the process by which team members identify and define the specific problem, to which they will apply the problem solving process. The tools and techniques presented in this section include: Brainstorming; Nominal Group Technique; Multivoting;

Problem Solving and Process Improvement Tools and Techniques

Page 22

Brainstorming What is Brainstorming?

Brainstorming is a technique for generating a large number of ideas in a short period of time. It is a group activity in which all members try to think of as many ideas as possible on a given topic as possible. The focus is on the quantity of ideas generated, not on the quality, as all ideas are evaluated later. Brainstorming assumes that no matter how crazy or irrelevant an idea may seem at first, it may spark an original and worthwhile idea that no one would have thought of otherwise. Brainstorming can serve to: document what the team already collectively knows; stimulate creative thinking; get everyone involved; bring the group closer together to work and think as a team.

When to use Brainstorming

Brainstorming is a particulary valuable tool to help teams identify problems to work on. It can also be used for identifying: possible causes; identify possible solutions; identify ways to implement solutions.

Problem Solving and Process Improvement Tools and Techniques

Page 23

How do you use Brainstorming?

Although the term "brainstorming" is often use generically to describe any idea generating activity involving a group of people, the formal brainstorming technique follows some specific rules. The brainstorming process is usually progressed in stages, these are: Preparation; Idea generation; Review; Evaluation; Follow-up.

Preparation 1. Determine the purpose and establish some guidelines. Determine the purpose of the brainstorming exercise. Make it clear to the group the type of ideas that are being sought and establish guidelines with respect to any areas to emphasise or ignore. 2. Set a time limit. Allow enough time for people to contribute their own ideas and to build on the ideas of others. Do not make the time too short. People often contribute their most valuable ideas towards the end of a session, when they have run out of ideas and have to force themselves to come up with more. 3. Distribute all data relevant to the purpose of the meeting before the meeting, to avoid wasting meeting time briefing members on the purpose. Idea Generation Stage 1. Encourage contributions from everyone. Encourage group members to contribute any idea no matter how improbable. Keep the pace reasonably quick so that group members do not have time to self-censor their contributions. 2. List all of the contributions as they are offered. This ensures that ideas do not get lost, it avoids duplication and enables other group members to build on other people's ideas. Often a completely impractical idea from one person will spark another member to think of a creative and practical idea. 3. Don't evaluate. Criticism or comments about someone's idea may slow down or even kill off further contribution. The purpose of brainstorming is to generate as many ideas as possible, so encourage all contributions.
Problem Solving and Process Improvement Tools and Techniques

Page 24

Some typical idea spurring questions to assist group members with building on each others ideas include: What else is like this ? What other ideas does this suggest ? Greater frequency ?

Review Review the list of idea, to ensure all team members understand what has been written down. This review is done purely as an aid to understanding and clarification. There is no discussion or criticism at this stage. Evaluation Evaluate all the contributions, to: combine or consolidate duplicate and related items; exclude items irrelevant to the topic; exclude personnel related items - we are focussing on the problem with the process;

Discuss and evaluate each item, if discussion generate new items, add these to the list. Conclude the session. Distribute copies of all the ideas generated to the team members after the session, for individual reflection. Follow-up Stage A follow-up team meeting is held to prioritise and reduce this list of ideas. This is usually done using the list reduction, multi-voting or nominal group techniques.

Problem Solving and Process Improvement Tools and Techniques

Page 25

Leading a Brainstorming Session

There are two general methods for leading a brainstorming session. The most familiar is the free wheeling method, where: Group members call out their ideas spontaneously A scribe records the ideas as they are offered

A variation on this is round-robin brainstorming, where: The leader asks each member in turn, for an idea Members may pass on any round The session continues until all members have passed during the round Ideas are recorded as in free wheeling

Problem Solving and Process Improvement Tools and Techniques

Page 26

List Reduction What is List Reduction?

List reduction is a technique that can be used to evaluate and assess the results of a brainstorming or other session that has generated an extensive list of ideas, so as to reduce it to a significant few that can be investigated in detail.

How to use List Reduction

The group identifies some criteria or "filters" that should be satisfied for an item to remain on the list. Some typical criteria might be: Is this item likely to improve the situation ? Is this feasible ? Is the cost excessive ?

Keeping the agreed upon criteria in mind, group members review or in appropriate cases vote on each item. Any borderline items should be bracketed or underlined rather than crossed, out so that group members may return to them later if necessary. In general, the group focuses on and continues to evaluate the items that remain on the list. The process may be repeated with different or more stringent criteria, until the list is reduced to a manageable number of options for applying some of the other analytical tools.

Problem Solving and Process Improvement Tools and Techniques

Page 27

Nominal Group Technique What is the Nominal Group Technique (NGT)?

The Nominal Group Technique is a structured, decision oriented form of brainstorming that equalises the contribution of each member during the ideas generation process and gives members equal voice during decision making. This overcomes the problem of the process being overly dominated by a few individuals that can occur with brainstorming. NGT is particularly suitable for situations requiring a degree of technical expertise, as is the case with the more complex problems.

When do you use the Nominal Group Technique?

NGT may be used in place of brainstorming or a method of prioritising the list of ideas generated through brainstorming and then reaching a decision on what to do next.

How do you use the Nominal Group Technique?

The Nominal Group Technique process is progressed in stages. These are: Idea generation; Idea presentation; Idea discussion; Evaluation; Prioritise; Make a decision.

Idea Generation The preparation procedure used for brainstorming should be also used for NGT. At the team meeting: the team leader or facilitator asks the team members to quietly generate ideas, on an individual level, for about 5-to-15 minutes; No discussion should be allowed and no-one is allowed to leave the room or in any other way distract other team members needing more time.

Problem Solving and Process Improvement Tools and Techniques

Page 28

Idea Presentation The idea presentation stage is progressed as follows: The facilitator asks each team member in turn to volunteer one idea. The ideas are written up on a flip-chart or white board, visible to the team members as they are offered. Members are encouraged to feed off each others ideas ie. if some one thinks of a new idea after hearing some one elses idea, they add it to their list and offer it when it is their turn next. This process continues until everyone has offered all their ideas. No discussion or criticism is allowed at this stage.

Idea Discussion Review the list of ideas, to ensure all team members understand what has been written down. This review is done purely as an aid to understanding and clarification. Again, no discussion or criticism at this stage. Evaluation Evaluate all the contributions, to: combine or consolidate duplicate and related items; exclude items irrelevant to the topic; exclude personnel related items - we are focussing on the problem with the process;

Discuss and evaluate each item, if discussion generate new items, add these to the list. Prioritise An approach to prioritisation used during NGT is as follows: Each member writes out their own copy of all the ideas left over from the previous stage; They individually rate their idea ie. give each idea a score out of ten, or allocate 100 points between the ideas etc.; The team leader or facilitator asks each member to call out their score each idea; The scores are totalled to give the overall rankings ie. the highest score being the highest ranked.

Problem Solving and Process Improvement Tools and Techniques

Page 29

Make a decision Where a decision needs to be made, so as to arrive at a decision on a single idea or option, the following extension to the above process can be applied. Take the ranked ideas which resulted from the prioritisation step and eliminate the lowest ranked item from the list; Reprioritise the items on the list using the prioritisation process previously described; Eliminate the lowest ranked idea; Repeat the process until only one item remains.

Alternately: Select the highest ranked item. Variation A variation to the basic ranking and decision making technique presented here is to: Define several relevant criteria against which to rate the ideas; Apply a rating for each criteria to each idea; Develop a consolidated score that encompasses all the criteria.

Problem Solving and Process Improvement Tools and Techniques

Page 30

Multivoting What is Multivoting?

Multivting is a technique used to reduce a list of ideas or problems down to a few that the team can focus on. It is similar to the prioritisation technique used in the nominal group process. However, in multivoting, the team never multivotes down to only a single item.

When do you use Multivoting?

Use multivoting after a brainstorming sessions, to prioritise and reduce the list ideas and problems team members have raised.

How to you use Multivoting?

1. List and number all the ideas. The team leader or session facilitator lists and numbers all of the ideas being considered for prioritisation. eg. on a flip chart. Ensure every one in the team can see them. 2. Take the first vote. Each team member votes on the item they feel has the highest priority. Team members can vote for as many or as few items as desired. 3. Elimination. Eliminate those items that received a relatively low number of votes. 4. List and renumber. List and renumber the remaining items. 5. Take the second vote. For the second vote, each team member is given a specific number of votes, equal to half of the number of items remaining on the list. Team members then allocate their votes to the remaining items. 6. Elimination. Eliminate items receiving relatively few votes. 7. Recycle the process. Continue voting using the procedure followed during the second vote, until only a few items remain. The remaining items are the ones which the team will focus their problem solving activities on.

Problem Solving and Process Improvement Tools and Techniques

Page 31

5. The Seven Tools of Quality

This section presents the seven basic tools of Quality. These are the original seven tools put forward by Dr. Kaoru Ishikawa. They are: Check sheets; Histograms; Cause-and-Effect diagrams; Pareto Charts; Scatter Diagrams; Run Charts; Control Charts.

The use of each of these tools shall now be discussed in detail.

Problem Solving and Process Improvement Tools and Techniques

Page 32

Cause and Effect Analysis (Fishbone Diagram) What is a Cause and Effect Analysis?
Cause and Effect analysis is a technique for identifying the possible causes of a know effect, usually the problem to be solved. The Cause and Effect diagram is also known as a "Fishbone diagram" or an Ishikawa diagram, after the Japanese Quality Management statistician Dr. Kaoru Ishikawa who developed the technique in 1943. The Cause-and-Effect diagram is a visually effective method of recording the possible causes of a problem being analysed as they are suggested, organised into the major categories that potentially contribute to a problem. It is effectively a structured form of brainstorming. The use of the fishbone diagram to provide structure to brainstorming activity provides a degree of focus on the problem at hand and more clearly allows team members to see or establish the link between the observed effect and possible causes.

When do you use Cause and Effect Analysis?

The fishbone diagram is usually used after the problem identification/selection stage of the problem solving process but before the data collection stage. Once a problem has been identified using one the techniques identied earlier, the fishbone diagram is used to focuss the team thinking on possible causes. The fishbone diagram will assist team members to determine what issues they should be collecting data on, during the data collection stage of the problem solving process. Another advantage of cause-and-effect diagrams is that one can tell at a glance how thoroughly a problem has been investigated. A diagram which contains a lot of detail would suggest that a group has delved deeply into a problem. At a more macro level, a fishbbone diagram can be used as an overall problem identification tool, by deliberately using a broad or vague problem statement to start the process eg. improve business performance, improve quality. The outcomes of the fishbone exercise will be some more specific problems on which improvement teams may focus.

Problem Solving and Process Improvement Tools and Techniques

Page 33

How do you build a Cause and Effect Diagram?

Constructing a Cause and Effect diagram is a four step process. This is illustrated on the following page in Figures 5.1.a to 5.1.d. 1. Describe the problem or "effect". Write a summary statement of the problem in a box on the right of the chart, with an line across the chart representing the backbone See Figure 5.1.a.
Installation Delays

Figure 5.1.a. Starting the Fishbone 2. Decide on cause categories, add these to the diagram. These major categories of causes are placed parallel to and some distance from the main process arrow. The major categories are then connected by arrows slanting towards the main arrow. The diagram will resemble a fishbone from which it gets its name. See Figure 5.1.b.

People

Procedures

Installation Delays

Equipment

Environment

Figure 5.1.b. Adding the Problem Category Spines What constitute meaningful categories may vary, depending on the nature of the problem. Unfortunately, when commencing a cause and effects analysis, we will often not be able to anticipate the problem categories which may emerge as we progress with the exercise. Therefore, we usually use some common generic categories to get the process started eg. common starters used in service businesses include:
Problem Solving and Process Improvement Tools and Techniques

Page 34

people; procedures; environment; equipment.

In manufacturing businesses, common starters are: people; methods; machines; materials.

After conducting our initial cause and effect analysis, we may find that the results obtained are good enough to proceed with data collection. Alternately, our first attempt at cause and effect analysis may potential causes naturally fall into certain categories. At this wish to redraw the fishbone diagram using categories that we specific to the particular problem being analysed and continue greater depth using these categories. reveal that the point we may now know are the analysis in

3. Brainstorm for possible causes. Write these on the chart clustered around the major category of causes they most influence. Causes should be divided and subdivided to show, as accurately as possible the various elements of each and how they interact. See Figure 5.1.c.

People
Sales people

Procedures
Poor inventory control

Technician Poor sales notification

Truck availability Computer problems

Installation Delays
Bad weather Heavy workloads Heavy traffic

Equipment

Environment

Figure 5.1.c. Building the Fishbone Diagram

Problem Solving and Process Improvement Tools and Techniques

Page 35

Where team members identify causes related to causes that have already been added to the fishbone diagram, these subsequent causes are shown as extensions to the ones already identified, as illustrated in Figure 5.1.d.

People
Sales people Wrong skills Not available too busy Not enough trucks Truck breakdowns Computer problems Back ups Truck availability No stock on hand

Procedures

Technician

Poor inventory control Poor sales notification Order not in system

Installation Delays

Bad weather Heavy workloads Heavy traffic

Co mpu ter d ow n

sh cr a r te pu m Co

Equipment

Environment

Figure 5.1.d. Continuing to build the Fishbone 4. Evaluate the possible causes. Look at each item on the fishbone and ask how it could contribute to the problem. Either list the responses next to the item or draw them as extensions to the fishbones as shown is Figure 5.1.d. This analysis should also reveal how the causes provoke the effects and assist in developing appropriate solutions.

Variations and Ideas

There are several methods for available for developing Cause and Effect or "fishbone" diagrams. The random method requires group members to suggest causes which apply to any of the major sub-divisions of the diagram. As in brainstorming, the session has a leader and a scribe to record the contributions of the group.

Problem Solving and Process Improvement Tools and Techniques

Page 36

The systematic method requires the leader to choose one of the major subdivisions on which to focus the group's attention. The brainstorming process looks at each sub-division in turn until the diagram has been completed. When constructing the fishbone diagram, attention to a few facilitation pointers will add value to the exercise: Use large charts and use large printing - ensure everyone can see what is going on;. To encourage a free flow of ideas, follow the usual brainstorming rules; Do not overload the diagram. If a group of causes begins to dominate, develop a separate diagram to explore this group in more detail; Construct a separate diagram for each problem. Where a problem definition is not specific enough, some major categories of the diagram may become overloaded and it will become necessary to redefine the problem; Look out for and examine closely the relationships between causes. This is where unexpected solutions are likely to turn up; Circle the most likely causes. This is usually done after all possible ideas have been posted on the fishbone diagram and each has been critically examined.

Preparing to Collect Data

In rare situations, an obvious implementable solution, which all team members agree on will emerge as a result of the cause and effect analsysis. However, the usual result of such a session is to go and collect further information in order to evaluate the key causes that emerged from the fishbone diagram. The results of the cause and effect analysis can be used to guide the data collection exercise and form the basis of the check sheet design. Developing the Check sheet To derive the required check sheet format from the fishbone analysis, the following steps can be followed: Evaluate the identified causes to identify potential root causes; Apply prioritisation, ranking and list reduction techniques to reduce the list and identify the most likely or key causes; Develop a checksheet which includes the reduced list of causes.

Problem Solving and Process Improvement Tools and Techniques

Page 37

Check Sheets What is a Check Sheet ?

A check sheet or tallysheet is a form that allows you to systematically collect data. You enter the data as check marks, ticks or measurements and analyse them later. Checksheets are usually used for recording numerical data. There is no standard form for checksheets. They should be set up to whatever form is easiest to record and analyse the information.

When to use a Checksheet

One type of checksheet is needed when you are recording data that you are counting, such as how often something happens over a given period of time, a different type is used for recording measurements. Checksheets are particularly useful when several people are collecting data. They ensure that everyone will collect comparable data in the same format, and they also provide a clear record of gathered data.

How to use a Checksheet

1. Determine the type of data that you need to collect. This should guide the user to the type of check sheet required. You will be trying to answer the questions: Will the data gathered reveal facts ? Can the data be analysed in such a way so as to reveal facts ?

Therefore, decide initially all the factors which might affect a situation and how you might analyse the information when you collect it. 2. Determine the categories or types of data possible. It will not be necessary to define all the categories at the start. The check sheet can be expanded in light of actual experience. 3. Design the checksheet form for people to use as they record the data. The sheet must reflect the type of data collected and reflect the categories of data that are likely to be expected. 4. Test the checksheet Get someone who didn't help design it to use it. Make any revisions that are necessary.

Problem Solving and Process Improvement Tools and Techniques

Page 38

Design a master checksheet if required to combine the results of numerous individual forms where several people will be involved in collecting the data. 5. Proceed to gather data. Sample checksheet for a counting example is shown in Table 5.1.

Category 1 2 3 4 5 6 7 8 Allergy Broken limb Back/Neck Concussion Contusion Heart Attack Sprains Other

Tally

Frequency 8 12 6 14 7 1 23 3 Total 78

Table 5.1 Causes of Lost Time Injuries Table 5.2 illustrates a measuring example. Table 5.2 Box Weight Sampling Weights Sample 2 509 501 500 507 508 in Grams Sample 3 499 505 502 504 498

Batch Number R1-710-4 R1-710-5 R1-710-6 R2-710-1 R2-710-2

Sample 1 502 497 501 500 502

Sample 4 501 502 499 503 499

Average 503 501 501 504 502

Problem Solving and Process Improvement Tools and Techniques

Page 39

Pareto Analysis What is Pareto Analysis?

Pareto analysis is a technique for recording and analysing information relating to a problem or a cause, which readily enables the most significant contributors to be identified. The resulting display, referred to as a Pareto Chart is a special form of histogram which allows the information to be visually displayed. A pattern will usually emerge when we look at the relationship between the numbers of items/occurrences of any situation and their relationship to the cause under consideration. This pattern has been referred as the "80/20" rule.

The Pareto Principle

This form of analysis and the resulting chart is named after an Italian economist Vilfredo Pareto, who studied the distribution of wealth in Europe in the late 18th and early 19th centuries. He concluded that 80% of the wealth was held be 20% of the population. Thus the Pareto Principle or 80/20 rule was born. During the 20th century Dr. J. Juran was able to show that this was a much more general principle as a result of his studies into business activities. For example: 80% of a company's waste are as a result of 20% of the causes; The bulk of a company's business comes from relatively few customers; 80% of your phone calls come from 20% of your friends and colleagues.

Although the exact ratio is not critical, the principle that problems can be classified as either being one of the vital few or the trivial many is important, as it provides considerably more focus to our problem solving efforts and yields disproportionately beneficial results. Pareto analysis shows at a glance which problems can be treated as the vital few, which are the trivial many and assists which the allocation of priorities. Although this technique will confirm many known notorious problems as belonging to the vital few, one of its major benefits is in being able to flush out problems not previously thought of as significant, identifying these as also belonging to the vital few.

Problem Solving and Process Improvement Tools and Techniques

Page 40

When do you use Pareto Analysis?

Pareto analysis can be used for several purposes. Namely, to: define where to concentrate problem solving efforts to get the most impact; display the relative importance of the various components of a problem; narrow a list of possible causes to identify the most likely cause; to identify root causes.

How to Construct a Pareto Chart

1. Determine the categories to be plotted on the chart. Typically you will want to categorise the data to be analysed. These categories will most likely be those against which data was collected on the check sheet used during the data collection stage. 2. Determine the measurement scale to use. The most common are cost, frequency of occurrence and percentage. 3. Collect and tabulate the information required. See the previous section on check sheets for more information. The following is an example of a record of machine stoppages recorded for a wrapping machine. Table 5.3. Machine Stoppages Causes Number of Occurences Machine breakdown 9 Operators Error 14 Wrapping Jam 62 Mechanical Jam 18 Product Jam 38 Foreign Object 3 Other 8 Total 152 Percentage 6% 9% 41% 12% 25% 2% 5% 100%

Problem Solving and Process Improvement Tools and Techniques

Page 41

4. List the categories along the bottom of the chart, starting with the largest one on the left. If there are many very small categories, group these together into a single group called "other". See Figure 5.2.a.

O p e r a t o r P r o d u c t W r a p p i n g M e c h . M a c h i n e F o r e i g n O t h e r E r r o r J a m J a m J a m B / d o w n O b j e c t
Figure 5.2.a. Categories Layout for Pareto Chart 5. On the left side, mark off the vertical axis representing the measure that was used. Mark off the units of measure up to or slightly more than the maximum occurrences measured for any one category.
7 0 6 0 5 0 4 0 3 0 2 0 1 0 0 O p e r a t o r P r o d u c t W r a p p i n g M e c h . M a c h i n e F o r e i g n O t h e r E r r o r J a m J a m J a m B / d o w n O b j e c t

Figure 5.2.b Mark up the Vertical Axis

Problem Solving and Process Improvement Tools and Techniques

Page 42

6. Draw a bar above each category to represent its size in terms of the measurement which was used. From Table 5.3, draw a bar respresenting the nubmer of occurences for each catergory of cause. This is illustrated in Figure 5.2.c.

7 0 6 2 6 0 5 0 4 0 3 0 2 0 1 0 1 8 1 4 9 8 3 0 O p e r a t o r P r o d u c t W r a p p i n g M e c h . M a c h i n e F o r e i g n O t h e r E r r o r J a m J a m J a m B / d o w n O b j e c t
Figure 5.2.c. The completed Pareto Chart 7. Interpret the Pareto Chart From Figure 5.2 we can see that wrapping jam is the major cause, followed by product jam. The remainder of the causes are relatively insignificant at this stage when compared to the first two. The Pareto chart suggests, that to eliminate the problem of wrapping machine stoppages, focussing our efforts initially on resolving the reasons why wrapping jams occur and then on product jams will give a disproportionately high improvement in performance, when compared to putting effort into the other causes. ie. These are the Pareto causes, accounting for some 66% of the problem. A word of caution! This is not to say that the other causes are not important and should be ignored. Improvement teams should work to eventually eliminate or minimise all the causes identified.

3 8

Problem Solving and Process Improvement Tools and Techniques

Page 43

However, what the Pareto chart tells us, is that if you want to make the biggest improvement in the shortest time or for the least effort, the Pareto causes should be addressed first. Only when these have been successfully resolved, should teams start addressing the other causes.

Variations and Enhancements

Pareto charts can be drawn using the absolute measures from the data collected as given in the second column of Table 5.3 and as illustrated by Figure 5.2.c. Alternately, the absolute measures can be converted to a percentage of the total, as has been done in third column of Table 5.3. This illustrated in Figure 5.2.d.

6 0 4 1 4 0 2 5

Percntag

2 0 1 2 9 6 5 2

0 W r a p p ' g M e c h . O p . F o r e i g n O t h e r P r o d u c t M a c h . J a m J a m E r r o r O b j e c t J a m B / d o w n
Figure 5.2.d. Percentages Pareto Chart 8. The information displayed on the chart can be enhanced by plotting a line showing the cumulative values as we move from left to right across the chart. The cumulative line on a Pareto chart is an indicator of the degree to which the problem is sensitive to the Pareto cause, or the degree to which the Pareto cause dominates the problem. ie. the steeper the cumulative line, the bigger will be the benefit of removing the Pareto cause. A percentage Pareto chart with cumulative line, for the data giving in Table 5.3 is shown in Figure 5.2.e.

Problem Solving and Process Improvement Tools and Techniques

Page 44

1 0 0

8 0

6 0 4 1 2 5

4 0

Percntage

2 0 1 2 9 2 0 W r a p p ' g M e c h .O p . M F o r e i g n t h e r P r o d u c t a c h .O J a m J a m E r r o r O b j e c t J a m B / d o w n 6 5

Figure 5.2.e. A Percentage Pareto with Cumulative Line

Problem Solving and Process Improvement Tools and Techniques

Page 45

Histograms What is a Histogram?

A histogram is a chart using bars of varying heights that illustrates the shape of a data distribution from which data has been collected.

When to Use a Histogram

Use a histogram when you need to obtain an understanding of the degree and nature of variations that are occurring in a given process, where they are occurring and to what extent.

How to Construct a Histogram

1. Define what you are going to measure Example: Age at failure of washing machines in customers homes. 2. Collect the data. The following represents data collected over a one month period. Age in months, of washing machines at failure 56 12 1 80 41 15 68 70 46 49 72 63 34 68 71 48 44 60 49 54 56 63 50 50 57 45 68 47 68 42 32 40 57 37 63 29 43 60 49 23 67 59 51 52 74 49 61 78 72 53

3. Determine the number of classes into which to divide the data. When looking at a large group of figures, they can more easily be displayed and understood if they are grouped for convenience. The intervals that define the groupings are usually called classes" or "class intervals" and the number of data readings falling into each of the classes are called the "frequencies". The following table provides a good "rule of thumb" for determining the number of classes or intervals to use.

Problem Solving and Process Improvement Tools and Techniques

Page 46

Table 5.4 Determining the number of Class Intervals Number of Data Values (N) Under 50 50 to 100 100 to 250 Over 250 Number of Classes (K) 5 to 7 6 to 10 7 to 12 10 to 20

In our example, the number of data values is 50 ie. N = 50. Therefore, from the above table, we shall select the number of class intervals as 7. ie. K = 7. 4. Determine the class interval size. By examining our original data, we can observe that the largest value (which we shall arbitrarily call L) is 80. ie. L = 80. Similarly, the smallest value present is 1 (which we shall arbitrarily call S) ie. S = 1. The class sizes can now be calculated as: L - S K = 80 - 1 = 7 79 7 = 11.28

Which we shall round off to 12. 5. Construct a Tally Sheet with the appropriate class boundaries defined. With reference to our previous calculations, we can construct a Tally Sheet (also referred to as a frequency diagram) and analyse our data in preparation for plotting a histogram. This is shown in Table 5.5. Table 5.5 Tally Chart of washing machine failures Class No Class Boundaries 1 1 to 12 2 13 to 24 3 25 to 36 4 37 to 48 5 49 to 60 6 61 to 72 7 73 to 84 Frequency Tally Frequency or Count II 2 II 2 III 3 I I I I I I I I 10 I I I I I I I I I I I I 17 II I I I I I I I I 13 III III 3 Total = 50

Problem Solving and Process Improvement Tools and Techniques

Page 47

As a crosscheck that you have constructed the tally chart correctly, the frequency column should total up the number of values in the original data. In the example given here, N = 50. 6. Construct the Histogram. Write the class interval values at the bottom of the graph. Write the frequency measures along the vertical axis.

2 0

1 5

1 0

0 1 1 2 2 5 3 6 4 9 6 0 7 3 8 4 1 3 2 4 3 7 4 8 6 1 7 2
Figure 5.3.a. Laying out the axes for a Histogram 7. Draw in bars representing number of items counted in each of the class intervals. The completed histogram is show in Figure 5.3.b. 2 0

1 7 1 5 1 3 1 0 1 0

5 2 2 3 3

0 1 1 2 2 5 3 6 4 9 6 0 7 3 8 4 1 3 2 4 3 7 4 8 6 1 7 2
Figure 5.3.b. Completed Histogram
Problem Solving and Process Improvement Tools and Techniques

Page 48

8. Evaluate the Histogram. Look at the histogram and assess: What shape is the histogram ? Is the distribution symmetrical ? Is it skewed ? How many peaks are there ? Are there any isolated bars ? What is the most common value ? Is it cliff like ? Does it look like a comb ?

In other words, what are the characteristics of the information gathered. Some examples of the many possible data shapes are illustrated in Figures 5.3. b. to 5.3.e.

Figure 5.3.b. A Skewed distribution (skewed right) A skewed or "cliff like" distribution such as this one suggests the possibility of preselection to remove data below certain values or that their exists some "natural" barrier below which no values are possible.

Figure 5.3.c. Comb like distribution

Problem Solving and Process Improvement Tools and Techniques

Page 49

A "comb like" histogram suggests poor measurement method, faulty or incorrectly set instrumentation or similar recording problem.

Figure 5.3.d Bi-modal distribution This histogram is "bimodal" ie. it has two data peaks. Bi or multi-modal histograms suggest that data from two or more populations has been mixed together, as is the case when there are two or more different processes at work, which the observer may have confused or assumed to be as one.

Figure 5.3.e. Outlying peak An outlying peak such as on this histogram like this is suspicious. Investigate the cause(s) of the peak on the far right as it appears to be an unusual occurrence when compared to the remainder of the distribution.

Problem Solving and Process Improvement Tools and Techniques

Page 50

The Scatter Diagram What is a Scatter Diagram?

Scatter diagrams, often called a scattergrams, show whether there is a relationship between two different characteristics of some process, operation or other phenomenon. Measurements are plotted as dots on a graph and the resulting pattern should indicate whether or not there is a relationship. eg. an increase in one element corresponds to an increase in the other etc.

When to Use Scatter Diagrams

Scatter diagrams are used to uncover relationships that could indicate possible causes. In particular, scatter diagrams can help establish whether an effect is related to a particular cause. Some examples could include, does working more overtime hours affect the numbers of work related injuries? Does high customer satisfaction lead to increased sales?

How to Construct a Scatter Diagram

1. Determine which two characteristics are to be studied Example: the amount of business generated is related to the number of calls made by sales representatives? 2. Collect the data. Collect paired samples of data. Typically at least 50 to 100 should be collected to ensure sufficient accuracy. For the purposes of this example, as lesser number shall be used. The data is shown in Table 5.6.

Problem Solving and Process Improvement Tools and Techniques

Page 51

Table 5.6 Sales data for last calendar year Month January February March April May June July August September October November December Sales Calls 7 8 18 44 20 24 30 32 12 18 12 4 Sales ($000) 300 350 1100 2000 1250 1300 1600 1650 800 900 750 200

3. In preparation for graphing, mark off measures for one characteristic on the horizontal axis and measures for the second on the vertical axis. Ensure the low end of both scales is in the lower left hand corner. This is illustrated in Figure 5.4.a.
2 , 5 0 0

2 , 0 0 0

1 , 5 0 0

Salesin$'0

1 , 0 0 0

5 0 0

0 0

1 0

2 0

3 0

4 0

5 0

N u m b e r o f S a l e s C a l l s

Figure 5.4.a. Laying out the axes for a Scatter Diagram 4. Plot the information on the graph. For the example given, for each month of the year, plot a dot, cross or other marker on the graph paper, positioning it in relation to both the horizontal and vertical axes. This is illustrated in Figure 5.4.b.

Problem Solving and Process Improvement Tools and Techniques

Page 52

2 , 5 0 0

2 , 0 0 0

1 , 5 0 0

Salesin$'0

1 , 0 0 0

5 0 0

0 0

1 0

2 0

3 0

4 0

5 0

N u m b e r o f S a l e s C a l l s Figure 5.1.b. The Completed Scatter Diagram

5. Evaluate the pattern and determine if a relationship exists. In some cases it may be possible to draw a line of best fit through the dots. This will show the expected trend of the data. Figure 5.4.b. shows a strong positive correlation between the number of sales calls and sales revenue. ie. the more calls that are made, the more the sales revenue. Examples of the types of correlations that can be observed with scatter diagrams are given in Figures 5.4.5.c. to 5.4.j.
1 2 0 1 0 0 8 0 6 0 4 0 2 0 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0

Figure 2.4.d. A strong negative correlation

Problem Solving and Process Improvement Tools and Techniques

Page 53

1 2 0 1 0 0 8 0 6 0 4 0 2 0 0 0

1 0 2 0 3 0 4 0 5 0 6 0 7 0

Figure 5.4.e. A weak negative correlation

1 0 0

8 0

6 0

4 0

2 0

0 0

1 0

2 0

3 0

4 0

5 0

6 0

7 0

Figure 5.4.f. Weak positive correlation

1 2 0 1 0 0 8 0 6 0 4 0 2 0 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0

Figure 5.4.g. No correlation

Problem Solving and Process Improvement Tools and Techniques

Page 54

1 0 0

8 0

6 0

4 0

2 0

0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0

Figure 5.4.h. Non-linear (hyperbolic) correlation

2 , 5 0 0

2 , 0 0 0

1 , 5 0 0

1 , 0 0 0

5 0 0

0 0

1 0

2 0

3 0

4 0

5 0

Figure 5.4.i. Stratification

1 0 0

8 0

6 0

4 0

2 0

0 0 1 02 03 04 05 06 07 0

Figure 5.4.j. Non-linear (Inverse Hyperbolic) correlation

Problem Solving and Process Improvement Tools and Techniques

Page 55

Some Notes about using Scatter Diagrams

When using the scatter diagram, note the following important points: Correlation does not mean that a cause and effect relationship exists between the two characteristics being plotted. It only means the two are related is some way, maybe through a third characteristic we are yet to test, it does not mean that one causes the other; If you are putting data from different sources onto a scatter diagram. it is possible that differences between the sources will interfere with the whole picture. Overcome this by using separate diagrams or by using different colours for data from different sources; Correlation may only exist over part of the range over which the data has been collected. Be aware so as not to misread a diagram and assume that correlation exists over the entire range; Negative relationships can be just as important as positive relationships.

Problem Solving and Process Improvement Tools and Techniques

Page 56

Run Charts What is a Run Chart ?

A Run chart is a graph in which a line is used to sequentially connect the data points. It gives a pictorial representation of what is happening to the process or system being measured, over time. It can also be used to show the relationship between two variables in which case it is referred to as a Line Graph. A Run Chart is easy to look at and any patterns in the data become readily apparent.

When to Use a Run Chart

Run charts are used when there is a need to monitor trends over time, or when the team suspects the cause of a problem is due to changes in a process or system occurring over a period of time.

How to Construct a Run Chart

1. Determine the process/system characteristic or measure to monitor. Example: A courier firm promising overnight parcel delivery has been experiencing increasing customer complaints about late deliveries. A run chart will be developed to monitor the number overnight delivery parcels delivered late each month. 2. Collect the required data. The number of late deliveries during the past year were extracted from the computerised logistics management system. These are given in Table 5.7. Table 5.7. Overnight parcel deliveries delivered late during past 12 months Month January February March April May June Number 160 157 169 131 145 120 Month July August September October November December Number 133 125 151 169 163 165

Problem Solving and Process Improvement Tools and Techniques

Page 57

3. Layout the axes for your chart. Draw the horizontal axis (x-axis) scaled by the appropriate time interval, in our example this will be months. Draw the vertical axis (y-axis) scaled by the frequency measure used. This is shown in Figure 5.5.a.
180

Number of Late Deliveries

170 160 150 140 130 120 110

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Month

Figure 5.5.a. Laying out the axes for a Run chart 4. Plot the data point for each month. Plot each individual data point and connect them up with a line. This is shown in Figure 5.5.b.
180

Number of Late Deliveries

170 160 150 140 130 120 110

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Month

Figure 5.5.b. Plotting the points on a Run chart 5. Calculate and plot the average (mean) of the data. From the data given in Table 5.7, calculate the mean or average number of late deliveries per month for the whole year. This has been calculated as 149.

Problem Solving and Process Improvement Tools and Techniques

Page 58

Draw a line representing the average on the run chart. This is also shown on Figure 5.5.c.
180

Number of Late Deliveries

170 160 150 149 140 130 120 110

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Month

Figure 5.5.c The completed Run chart

A Note on Interpreting Run Charts

A "normal" Run Chart should show a relatively equal distribution of points above and below the average line. Patterns which shown wide variations may indicate problems with an unstable process, a "slack" process or merely indicate time related cycles present in the system such as seasonal variations. Run charts may be used to show several groups of data simultaneously. Each group may be plotted as a separate line on the same graph using different symbols to mark the points. This may allow relationships between data to become apparent where they are present, eg. trends in one measure are reflected by a trend in the other measure, and reflect correlations between measures.

Problem Solving and Process Improvement Tools and Techniques

Page 59

Control Charts What are Control Charts?

A Control Chart is a graphic comparison of how a process is performing with how the process should perform. It is basically a run chart with some additional information added, in the form of lines representing control limits for the process being monitored. Control limits are statistically determined upper and lower boundaries that define the range of measurements that would be considered to be normal for a given process. A Control Chart enables you to distinguish between variation that is a natural part of a process, and that caused by something outside the process that could indicate a problem.

The Control Chart Principle

The principles on which control charts are based are discussed in more detail in Appendix 1 - Understanding Variation and Data. To briefly summarise some of the key points: Control charts use the knowledge of the standard deviation of data around its mean, which has been collected about a process, to determine what the natural limits of performance for that process are; The standard deviation and mean can be calculated for any set of data; For data which follows the Normal distribution, the range defined between the mean minus three times the standard deviation (called the lower control limit) and the mean plus three times the standard deviation (called the upper control limit), will enclose 99.73% of all possible data points that can be obtained from such a process; The upper and lower control limits will therefore indicate the natuaral limits of variation of a process. ie. 99.73% of the time, the process will yield a data point within these limits; Although not all data follows the Normal distribition, most data collected from real processes are close to normally distributed, and where they are not, the sampling plan used to collect the data can yield normally distributed points.

Problem Solving and Process Improvement Tools and Techniques

Page 60

When to use Control Charts

Control Charts can be used to address two broad categories of problems: Process analysis involves understanding the nature of variation in processes by looking at the effect of different process factors, so as to identify potential problems; Process control involves monitoring a process for which has been standardised, for which control limits have been set, to determine if the standardisation was correct and if it is being maintained.

Process control and process analysis charts are made in the same way, except their purposes are different.

Types of Data and Types of Charts

The form of Control Chart used will depend on the type of data that is to be plotted. Data will be of two basic types: things that you measure, or; things that you count.

For things that you measure, such as measurements of some physical dimension (eg. length in millimetres), time (eg. waiting time in minutes), weight (eg. weight in kilograms) or other unit of measurement, data can take any value depending on its measure. Such data can take on what are known to as "indiscrete" values or variables. ie. they can be anything and they can vary from item to item. Measured or variable data is referred to as Continuous data. For things that you count, for example, the number of defective items in a sample, or the number of errors per form, the data can only take on certain or "discrete" values. This type of data is called Attributes data. For attributes data, their can be two types - categorical data and occurrence data. The first type of attributes data is where we decide that we only want to know whether an "attribute" is present or not. For example, we may be monitoring the number of claim forms that have been incorrectly filled out. Any error present on the form qualifies the form as
Problem Solving and Process Improvement Tools and Techniques

Page 61

"incorrectly filled out" irrespective of how many errors have been made on any particular form. It is either correct or not correct. Other examples of this type of data is typically of the form OK or Not OK, Pass or Fail, True or False etc. This type of data is called Categorical Data. The other type of "attribute" data is where we can have a count of more than one attribute per unit examined. For example, we may be interested in how many errors were made per claim form, not just whether the form was correct or incorrect. In this case we count the number of errors per form and can end up with a data count that is greater than the number of items examined. eg. if we examine 10 forms and find that there are an average of 3 errors per form, we can end up with 30 total errors for 10 forms. Such data is referred to as Occurrence Data. Different Control Charts are used for each of these different types of data, and in the case of attributes data depending on whether the sample size, also referred to as subgroup size, can be kept constant from measurement to measurement. This is summarised in Table 5.8. Table 5.8. Summary of Control Chart Types.

Continous Categorical Occurrence Data Data Data Lots of data X Sample size np chart c chart a n d R available chart constant (pn chart) Limited data i and mr Sample size p chart u chart available chart not constant

How to Construct a Control Chart

The following is a general procedure for constructing a Control chart. The specific steps to be used in constructing the various charts given in Table 5.8. will be presented in subsequent sections. 1. Select the process is to be monitored and determine what data will be collected for it. 2. Determine which type of chart is suitable. Once the data to be collected, or characteristic to be monitored has been established, determine which chart is applicable to that type of data.

Problem Solving and Process Improvement Tools and Techniques

Page 62

Refer to Table 5.8. or refer to the decision tree given in Figure 5.6 to decide which chart is appropriate. 3. Decide how often data is to be collected. The more frequently the event you wish to monitor occurs, the more frequently you will need to take measurements. For example: Measure queue outpatient waiting times at 10:00am and 2:00pm, measure manufactured pin diameters every hour. etc. The time interval between successive measurements or data collection is referred to as the sampling interval. 4. Decide how many data points to record each time. That is, how many measurements will you take or how much data will you collect each time measurements are take or data collected. This is the sample or subgroup size. eg. measure waiting times for first five customers, measure first 10 pins manufactured each hour. 5. Collect the data. Implement you sampling/data collection plan. During each sampling interval or at the predetermined time, collect one sample of data values. 6. Calculate the statistics of interest. Use the prescribed statistical formulas for the particular chart you have chosen to calculate the position of the: Centre Line or Mean (CL) Upper Control Limit (UCL) Lower Control Limit (LCL)

Problem Solving and Process Improvement Tools and Techniques

Page 63

N o tm u c h d a t a a v a i l a b l e M e a s u r i n g A r e y o u c o u n t i n g o r m e a s u r i n g ? C o u n t i n g N o C a t e g o r i c a l D a t a C o n t i n u o u s D a t a

ia n d m r C h a r t s

L o t s o f d a t a X b a r ( X ) a v a i l a b l e a n d R C h a r t s

C o n s t a n t S a m p l e S i z e n p C h a r t S a m p l e s i z e n o t c o n s t a n t p C h a r t C o n s t a n t S a m p l e S i z e c C h a r t S a m p l e s i z e n o t c o n s t a n t u C h a r t

C a n y o u h a v e m o r e t h a n o n e c o u n tp e r u n i t Y e s O c c u r e n c e D a t a

Figure 5.6 Control chart selection logic tree 7. Draw the control chart This involves laying out the chart format, and then plotting the data points. To lay out the chart format: Draw and scale the vertical and horizontal axes; Draw the Centre Line, Upper and Lower Control Limits.

Then plot the data values collected on the Control Chart 8. Analyse the Control Chart. In general, if all the points plotted fall within the control limits, the process can be said to be in control and operating normally. The existence of points outside the control limits or the existed or certain characteristic trends would suggest that the process is not in control and that special causes are present.

Problem Solving and Process Improvement Tools and Techniques

Page 64

The X Chart
The X-chart is the most basic of control charts. It provides a plot of all of the individual points measured. The X-chart provides a direct respresentation of the process.

X Chart Example
Consider a company which manufacturers and delivers product to meet specific customer orders. The days taken to fill the past 25 orders have been recorded and are given in the following table. Table 5.1 Number of days to fill customer orders Order Number 1 2 3 4 5 6 7 8 9 10 11 12 13 Days to Deliver 15 18 23 22 18 15 31 27 13 33 35 31 24 Order Number 14 15 16 17 18 19 20 21 22 23 24 25 Days to Deliver 29 26 17 14 12 14 30 25 24 20 15 18

Total 1. Calculate the mean value

549

_ Calculate the mean value, where the mean ( X ) is calculated as: _ X = Xn n Where:

Xn
n

= =

represents the sum of all the data points or values the number of values recorded

Problem Solving and Process Improvement Tools and Techniques

Page 65

In our example: X = 15 + 18 + 23 + ...... + 18 25 549 25 21.96 = 22.

2. Calculate the standard deviation The standard deviation, usually referred to by the symbol or the letters SD or S, is calculated using the equation:

_ = (Xn - X)2
n-1

These symbols have been previously defined. In our example, this is calculated as follows:

(15-22)2 + (18-22)2 + (23-22)2 + ...... + (18-22)2 24 1156.96 24 6.94 = = 48.21

6.9

3. Calculate the Upper and Lower Control Limits. The general formulas for the calculation of Upper and Lower Control Limits are:

_
Upper Control Limit (UCL) = X + 3 Lower Control Limit (LCL) = X - 3 For example, substituting our value of standard deviation, we get: UCL = = LCL = 22 + 3 X 6.9 22 + 20.7 22 - 20.7 = = 42.7 1.3

Problem Solving and Process Improvement Tools and Techniques

Page 66

4. Draw the chart This is shown in Figure 5.7.

5 0 U C L = 4 2 . 7 4 0

3 0 M e a n X = 2 2 2 0

1 0 L C L = 1 . 3 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5

Figure 5.7 X-Chart of Delivery Performance 5. Interpret the chart. The control chart suggests that the process is in statistical control, centred on a mean performance level of 22 days. The natural limits of variation for this process are between 1.3 and 42.7 days.

Problem Solving and Process Improvement Tools and Techniques

Page 67

The X-bar and R Chart _

An X-bar and R chart is one that shows both the mean value X (called X bar) and the range, R. It is the most common type of control chart for continuous data. The X-bar portion of the chart mainly shows changes in the mean value of the process (the location), while the R portion shown any changes in the dispersion or spread of the process.

X-bar and R Chart Example

An industrial equipment manufacturer and retailer is contemplating launching a new marketing initiative using their superior delivery performance as a source of differentiation from their competitors. Among several options, they are considering giving customers a rock solid guarantee that their order will be delivered within a predetermined number of days and offering a substantial cash rebate if an order is delivered late. In order to determine what they can and cannot guarantee, they have decided to monitor the performance of their existing order fulfilment process. When collecting data for X-bar and R charts, the data needs to be collected in the form of subgroups or samples. Ideally, each sample should include at least 2 data items but no more than 10. The optimum is approximately 4 or 5. In this example, we shall be tracking the delivery performance of 5 randomly selected orders each day, for a period of 10 days. The results are shown in Table 5.9. Table 5.9 Delivery performance in days from receipt of order Day (Sample) 1 2 3 4 5 6 7 8 9 10 Order A Order B Order C Order D Order E Average 11 4 4 5 4 4 6 2 9 2 9 2 1 2 7 4 7 4 2 4 6 7 4 3 3 5 3 3 1 3 4 2 3 6 4 4 4 2 2 10 2 4 3 2 5 9 3 5 7 6 6.4 3.8 3 3.6 4.6 5.2 4.6 3.2 4.2 5

Problem Solving and Process Improvement Tools and Techniques

Page 68

With reference to Table 5.9, we have sampled 5 orders each time we have taken a sample ie. 5/day. The subgroup size, usually referred to as n, is therefore 5. (ie. 5 orders).

Calculate the statistics of interest. _

1. Determine the location of the centre-line for the X Chart. Calculate the average waiting time for the period being monitored. For the above example this is the sum of all the daily averages divided by 10 (10 days).

=
Average Mean X

=
Average Mean X

= 6.4 + 3.8 + 3 + 3.6 + 4.6 + 5.2 + 4.6 + 3.2 + 4.2 + 5 10 = 4.4 Days

_
The daily averages are referred to as X. The average of the averages which defines the centre-line is called X double bar. 2. Determine the centre-line for the Ranges chart. To determine the centreline of the ranges chart, we must first determine the range of each group of data collected. The range is the difference between the largest and smallest value in each sample. eg. for the orders sampled during day one, the fastest was filled in 2 days, the slowest in 11 days. The range is therefore 11 days - 2 days = 9 days. The ranges for each of the samples are given in Table 5.10. Table 5.10 Ranges of order fulfillment samples. Day Range 1 9 2 5 3 3 4 4 5 4 6 5 7 4 8 3 9 8 10 8

Problem Solving and Process Improvement Tools and Techniques

Page 69

The range is referred to as R. Calculate the average range for the period: _ Average Range R = 9+5+3+4+4+5+4+3+8+8 _ 10 Average Range R = 5.3 Days This defines the centre-line for the range chart.

_
3. Calculate the Upper and Lower Control Limits for the X Chart. The formulas for the upper and lower control limits are as follows: = _ Upper Control Limit (UCL) = X + A2R = _ Lower Control Limit (LCL) = X A2R Where the values of the coefficient A2 may be obtained from Table 5.11 Table 5.11 Coefficients for X and R Charts _ X chart R chart LCL Sub-group size A2 D3 (n) 2 2.66 0 3 1.023 0 4 0.729 0 5 0.577 0 6 0.483 0 7 0.419 0.076 8 0.373 0.136 9 0.337 0.184 10 0.308 0.223

R chart UCL D4 3.267 2.574 2.282 2.114 2.004 1.924 1.864 1.816 1.777

Problem Solving and Process Improvement Tools and Techniques

Page 70

For the example used here, we sampled 5 orders each day, therefore the subgroup size n = 5. From Table 5.11. A2 = 0.577. Substituting into the above equations: UCL LCL = = 4.4 4.4 + 0.577 x 5.3 0.577 x 5.3 = = 7.5 1.3

4. Calculate the Upper and Lower Control Limits for the R Chart. The equations for the upper and lower control limits for the range chart are: _ _ UCL = D4R and LCL = D3R From Table 5.11, D4 = 2.114 and D3 = 0. The Upper and Lower Control Limits can therefore be calculated as: UCL LCL = = 2.114 x 5.3 0 = 11.2

5. Lay out the X-bar and R chart format. Draw the centre-lines, upper and lower control limits for the X chart on one graph and the centre-line, upper and lower control limits for the R chart on another. 6. Plot the values. Plot the daily averages (means) on the X chart and the values of the daily ranges on the R chart. See Figure 5.8.a. (X-bar chart) and 5.8.b. (R-chart).
1 0

X C h a r t
U C L = 7 . 5

6 M e a n X = 4 . 4 4

DaystoFilOrde

L C L = 1 . 3

0 1

9 1 0

S a m p l e D a y s

Figure 5.8.a The X-bar Chart

Problem Solving and Process Improvement Tools and Techniques

Page 71

R C h a r t
1 4 1 2 1 0 8

DaystoFilOrde

6 4 2 0 1 2 3 4 5 6 7 8 9 1 0

S a m p l e D a y s

Figure 5.8.b. The R Chart 7. Analyse the charts. In general, a normal control chart should have points evenly distributed between the control limits. This indicates that variation is occurring, but that it is within the normal limits for the process. Changes in the mean of the process will bring about changes in the X Chart. The R Chart will remain unchanged. Changes in the spread of a process will effect both the R Chart and the X Chart. Increases in spread will cause points on the R Chart to increase and points on the X Chart to show a greater spread and possibly go beyond the control limits. When interpreting X and R charts, the R chart should always be read first. The R chart shows process capability, it shows the degree of variation in the process and the degree of variation due to common causes. If the R chart is out of control, it is pointless to attempt to make any process adjustments on the basis of the X chart. If the R chart looks stable and the X chart is not, this suggests that the process is probably inherently stable, and that incorrect or unnecessary adjustment or some other factor is causing the X chart instability.

Problem Solving and Process Improvement Tools and Techniques

Page 72

The i and mr Chart

When there are few data points and/or data is only available at irregular intervals, we may not be able to collect enough points to construct X-bar and R-charts. For example: a process that runs in single batches with a long cycle time; situations where it is uneconomical to take multiple samples; where destructive testing is used to take measurements; etc., etc.;

In such situations the i - individuals, and mr - moving range charts are used. These charts are similar to the X-bar and R-charts except that they use individual points.

An i and mr Chart Example

A mining company operates a process for extracting small amounts of previously unrecoverable precious metals for waste slurry. The slurry is stored in tanks as it is generated, every few days when enough slurry has been accumulated, the batch is processed. The number of grams recovered from each batch is recorded every time the recovery process is run. The results of the past 25 runs is shown in Table 5.12.

Problem Solving and Process Improvement Tools and Techniques

Page 73

Table 5.12 Precious Metals recoveries during last 25 runs Run 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Total Grams recovered 15 18 23 22 18 15 24 27 13 33 32 29 38 29 26 17 11 13 14 30 25 24 20 15 18 549 Moving Range 0 3 5 1 4 3 9 3 14 17 1 3 9 9 3 9 6 2 1 16 5 1 4 5 3 136

Problem Solving and Process Improvement Tools and Techniques

Page 74

1. Calculate the moving ranges. This has already been done, and is given in Table 5.12. To calculate the moving range values, subtract the previous value of grams recovered from the current value. If the result is a negative number, record it as a positive number ie. the magnitude of the difference, ignoring the sign. The moving range value for the first number is recorded as a zero, because there is no previous number to subtract from it. 2. Calculate the mean of the moving ranges. The mean of the moving ranges is calculated using the formula: _ R = R1 + R2 + ..... + Rn-1 n-1 Where, n is the number of values recorded. Note: the sum of the range values is divided by n-1 because the first one will always be zero and therefore be ignored. In our example, the average moving range is: _ R = 3 + 5 + 1 + ..... + 3 24 = 136 24 5.7

3. Calculate the control limits for the mr-chart. As for the R-chart in the X-bar and R-chart pairing, the Upper and Lower Control Limits for the mr-chart are calculated using the same equations. Namely: _ LCL = D3R _ UCL = D4R Because our subgroup size will always be 2 for an mr chart, D3 and our LCL will always be equal to zero.

Problem Solving and Process Improvement Tools and Techniques

Page 75

Similarly, D4 will always equal 3.267 (ie. D4 for n=2 is 3.267). Therefore: _ UCL = 3.267 x R Which in our case is: UCL = = 3.267 x 5.7 18.6

4. Draw the mr-chart

2 0 U C L = 1 8 . 6 1 5

1 0 M e a n R = 5 . 7 5 L C L = 0 0 2468 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 13579 1 1 1 3 1 5 1 7 1 9 2 1 2 3

Figure 5.9.a. The moving range chart (mr-chart) _ If the moving range chart is in control, the mean moving range R may be used to calculate the UCL and LCL for the i chart. If the chart is not in control, remove the values contributing to the outliers from the source data. 5. Calculate the value of the centre line for the i - chart The average of all the values defines the centre line for the i chart. This is calculated as : _ X = Xn n _ X = X1 + X2 + X3 + ....... Xn n
Problem Solving and Process Improvement Tools and Techniques

Page 76

Which in our example is: _ X = 15 + 18 + 23 + 22 + .... + 18 25 = 549 25 21.96 = 22

6. Calculate the Upper and Lower Control Limits These are calculated using the equations used for the X-bar chart. That is: _ _ UCL = X + A2R _ _ LCL = X A2R The value of A2 is 2.66. Therefore: _ _ UCL = X + 2.66 R, and _ _ LCL = X 2.66 R In our example, X = 22 and R = 5.7. Substituting gives: UCL = = = LCL = = 22 22 + + 2.66 x 5.7 15.16 37.2 15.15 6.9

37.16 = 22 6.84 =

Problem Solving and Process Improvement Tools and Techniques

Page 77

7. Draw the i - chart

4 0 U C L = 3 7 . 2 3 5 3 0 2 5 M e a n X = 2 2 2 0 1 5 1 0 L C L = 6 . 9 5 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5

Figure 5.9b. The completed individuals chart (i-chart) 8. Interpret the chart One point appears to be due to a special cause. However, given that this is above the control limit, indicating a higher than expected recovery rate, it warrants investigation to see if the process can be improved so that the mean is shifted to this higher level.

Problem Solving and Process Improvement Tools and Techniques

Page 78

The np Chart.
The np chart or Number Defective chart, also referred to as the pn chart is a chart that shows the number of defective items produced by a process. np charts are used with Categorical data, ie. attribute data where values are measured in terms of OK or Not OK. The np Chart is based on the assumption that count data, where we are counting OK/Not OK tend to follow the Binomial distribution, which approximates the Normal distribution. As subgroup size increases, the accuracy of this approximation increases. One restriction applying to the use of the np chart is that it requires sample or subgroup size to be constant. ie. each sample has the same number of items in it.

np Chart Example
An automatic packaging line which fills and seals bulk wine containers for export, traditionally has damaged a certain percentage of containers during the process. These are scrapped and the wine is recycled. The extent of the problem is going to be investigated using an np Chart. The first 200 containers filled each hour were examined and the results recorded for 25 consecutive production hours. The collected data is given in Table 5.12.

Problem Solving and Process Improvement Tools and Techniques

Page 79

Table 5.12 Defective wine container data Sample (Subgroup) Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Total Sample (Subgroup) size 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 200 5,000 Number defective 15 18 23 22 18 15 44 47 13 33 42 46 38 29 26 17 5 7 14 36 25 24 20 15 18 610

The Subgroup size indicates how many containers were sampled each time a sample was taken. In our case, this was 200 and remained constant through out the data collection period. The number of defectives indicates how many of these were defective. Note: Subgroup size n should be greater than 50, and the expected mean value of defectives for each subgroup should ideally range from 3 to 4. To calculate the required statistics for an np chart we proceed as follows.

Problem Solving and Process Improvement Tools and Techniques

Page 80

1. First, calculate the average number defective. This calculation is to work out the average number defectives found per sample, overall. In our case, we took 25 samples and found 610 defective items. The average number of defectives is calculated as: _ np = total number of defectives = 610 number of samples 25 _ np = 24.4 2. Calculate the average fraction or proportion defective for all the data collected. Average fraction or proportion defective is calculated as follows: _ p = Total defective = np Total inspected n for the data in this example, _ p = 610 = 0.122 = 5,000

12.2%

This defines the centre-line of the control chart. Given that the sample size is constant, this could also have been calculated as: _ _ p = average number defective = np sample size n = 24.4 200 0.122 = 12.2%

3. Calculate the Upper and Lower Control Limits. The equations for calculating the Upper and Lower Control Limits for an np chart are: _ _ _ UCL = np + 3 x np(1-p) _ _ _ LCL = np - 3 x np(1-p) UCL LCL = = 24.4 24.4 + 13.9 13.9 = = 38.3 10.5

Problem Solving and Process Improvement Tools and Techniques

Page 81

4. Layout the control chart. Draw in the centre line (np) and the upper and lower control limits on your control chart proforma. Plot your data points on the control chart. The completed control chart is shown in Figure 5.7.

n p c h a r t
5 0

4 0 U C L = 3 8 . 3

3 0 M e a n n p = 2 4 . 4

Numberdfective

2 0

1 0

L C L = 1 0 . 5

0 13579 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5

S a m p l e H o u r s
Figure 5.10.a. The completed np chart Note, the np chart shows some points lying outside the control limits. These have been highlighted in Figure 5.10.b.

Problem Solving and Process Improvement Tools and Techniques

Page 82

n p c h a r t
5 0

4 0 U C L = 3 8 . 3

3 0 M e a n n p = 2 4 . 4

Numberdfective

2 0

1 0

L C L = 1 0 . 5

0 13579 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5

S a m p l e H o u r s
Figure 5.10.b. np chart with special causes highlighted Data points outside the control limits indicate that there are factors present and impacting on the process that are not a normal part of the process. These factors are special causes. A process with numerous points outside the control limits - known as outliers, is said to be out of control, or not in statistical control. Investigating and resolving the specific causes of the outliers become priority projects for quality and process improvement teams.

Problem Solving and Process Improvement Tools and Techniques

Page 83

The p Chart
The p Chart or Proportion Defective Chart, is a chart that shows the fraction of defective items produced by a process. As for np charts, p charts are used with Categorical data, ie. attribute data where values are measured in terms of OK or Not OK. The p Chart is based on the assumption that count data, where we are counting OK/Not OK tend to follow the Binomial distribution, which approximates the Normal distribution. As subgroup size increases, the accuracy of this approximation increases. The key difference between the np chart and the p chart is that it does not require sample or subgroup size to be constant, although it can still be used in cases where sample size is constant. This gives it several advantages over the np chart for monitoring categorical data. In some cases it may not be possible to establish a constant sample size eg. 100% sampling of market driven process - the sample size will probably differ how we structure our sampling plan. In other cases, we may be using a sampling plan based on taking a statistically significant smaller sample of a high volume process. As is common in high volume production environments. In such cases, if production volumes increase substantially over time, we may find ourselves having to increase our sample size in order to still have a statistically valid subgroup size. If sample size changes, before and after np charts will not be directly comparable. p charts will remain directly comparable regardless of what changes are made to sample size over time.

p Chart Example
With reference to the automatic packaging line which was used as the basis of the np chart example, suppose we change our sampling so that we record all the containers filled per hour, for 25 consecutive production hours. What we will find, is the number of containers filled each hour varies depending a number of factors, such as equipment problems, material availability problems, operator availability etc., giving us a sample size that is no longer constant, but which reflects what was actually made. The new data is shown in Table 5.13.

Problem Solving and Process Improvement Tools and Techniques

Page 84

Table 5.12 Defective wine container data Sample (Subgroup) Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Total Sample (Subgroup) size 115 220 210 220 220 255 440 365 255 300 280 330 320 225 290 170 65 100 135 280 250 220 220 220 220 5,925 Number defective 15 18 23 22 18 15 44 47 13 33 42 46 38 29 26 17 5 7 14 36 25 24 20 15 18 610 % Defect. p (%) 13.0 8.2 10.9 10.0 8.2 5.8 10.0 12.9 5.1 11.0 14.6 13.9 11.9 12.9 8.9 10 7.7 7.0 10.4 12.8 10.0 10.9 9.1 6.8 8.2 UCL (%) 18.8 16.5 16.6 16.5 16.5 16.0 14.6 15.1 16.0 15.6 15.8 15.3 16.5 16.4 15.7 17.3 21.6 19.4 18.2 15.8 16.1 16.5 16.5 16.5 16.5 LCL (%) 1.8 4.1 4.0 4.1 4.1 4.6 6.0 5.5 4.6 5.0 4.8 5.3 4.1 4.2 4.9 3.3 0 1.2 2.4 4.8 4.5 4.1 4.1 4.1 4.1

Problem Solving and Process Improvement Tools and Techniques

Page 85

To calculate the required statistics for a p chart we proceed as follows. 1. Calculate the average fraction or proportion defective for all the data collected. Average fraction or proportion defective is calculated as follows: _ p = Total defective = np Total inspected n for the data in this example, _ p = 610 = 0.103 = 5,925

10.3%

This defines the centre-line of the control chart. 2. Calculate the Upper and Lower Control Limits. The equations used to calculate the upper and lower control limits for a p chart are as follows. _ _ _ UCL = p + 3 x p(1-p) n _ _ _ LCL = p - 3 x p(1-p) n Using our previously calculated value for p = 0.013 gives: UCL = 0.103 + 3 x 0.304 n 0.304 n

LCL

0.103 -

A note on calculating the control limits for p charts. The control limit formulas use the sample size n, as part of the calculation. Because our sample size is different for each sample, we will need to calculate the UCL and LCL separately, for each individual point that is plotted. This has been done and is displayed in the right-most two columns on Table 5.12. If the subgroup size is constant, then the UCL and LCL will also be constant. Figure 5.11 shows the p chart for the data in Table 5.12.

Problem Solving and Process Improvement Tools and Techniques

Page 86

2 5 2 0 1 5

U p p e r C o n t r o l L i m i t

M e a n

PercntDefctiv

1 0 5 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 2 3 2 4 2 5

L o w e r C o n t r o l L i m i t

S a m p l e N u m b e r
Figure 5.11 Completed p chart

Problem Solving and Process Improvement Tools and Techniques

Page 87

The c Chart
The c - chart or non-conformities chart, is a chart that shows the number of occurrences of some event ie. a non-conformity per sampling period or interval. The c-chart is used to monitor Occurrence Data. For the p and np charts (Categorical data) we know both the number of defectives as well as the number of remaining (nondefectives) units in the sample. This is not always possible. In many situations we may be able to measure the number of non-conformities or occurrences of some event only. For example, if a retailer were monitoring customer complaints (the non-conformities) it may be impractical to try to measure the total number of customers which come into the store, so as to determine the total sample size. ie. they can easily measure the number of complaints per day, but the total number of customers per day is not known. In such cases, a c-chart is used. A c-chart is used when the sampling period, interval or unit is constant. The sampling unit may be a fixed length, area, quantity, time etc. Examples of fixed sampling units are: Complaints per day; Scratches per car; Errors per form; etc.

A key difference between Occurrence data and Categorical data, is that for categorical data, we cannot obtain a count of the parameter we are monitoring which is greater than the number of items in the sample. ie. if we sample 100 forms to monitor the number or percent defective (ie. incorrectly filled out) the most we could conceivably measure is 100 assuming they were all defective. For Occurrence data, we may obtain a measure which is greater than the number sampled. For example, if we a measuring the number of errors per form (Occurrence data), and there is an average of 3 errors per form, we will obtain a measurement of 300 which is significantly greater than the sample size.

Problem Solving and Process Improvement Tools and Techniques

Page 88

c - Chart Example
The method for constructing a c-chart will now be illustrated by way of an example. Consider a textile firm manufacturing plain white fabric which is later dyed to meet specific customer orders. Weaving errors and stains (usually oil) which often find there way into the fabric during the process cannot be allowed in the finished product which goes to dying. The dye highlights flaws and stains, causing such fabric to be rejected. To minimise the amount of finished product which is scrapped, the plain white fabric is inspected. Any oil or other stains are manually cleaned during inspection, any sections with weaving faults are highlighted to be cut out the main roll before dying. The rolls of plan white fabric are all prepared in 3,000 metre rolls. Information is collected by the fabric inspectors on the number of weaving faults found. This information is used by the weaving operators and mechanics to try to improve the weaving process. Data collected from 25 rolls is shown in Table 5.14. Table 5.14 Number of weaving faults found per roll Roll Number 1 2 3 4 5 6 7 8 9 10 11 12 13 Number of Weaving Faults 15 18 23 22 18 15 44 47 13 33 42 46 38 Roll Number 14 15 16 17 18 19 20 21 22 23 24 25 Number of Weaving Faults 29 26 17 5 7 14 36 25 24 20 15 18

Total

610

Problem Solving and Process Improvement Tools and Techniques

Page 89

1. Calculate the average number of non-conformities Calculate the average number of weaving faults for the 25 rolls for which data has been collected. _ c = X1 + X2 + X3 + ...... + X25 number of samples = 15 + 18 + 23 + ..... + 18 25 610 25 24.4

This defines the centre line of the control chart. 2. Calculate the Upper and Lower Control Limits. For occurrence data, the standard deviation is calculated as the square root of the average. That is: _ Standard Deviation = c The Upper Control Limit is calculated as: _ _ UCL = c+3c = = = = 24.4 + 3 24.4 24.4 + 3 x 4.94 24.4 + 14.8 39.2

The Lower Control Limit is calculated as: _ _ LCL = c-3c = = = 24.4 - 3 24.4 24.4 - 14.8 9.6

Problem Solving and Process Improvement Tools and Techniques

Page 90

3. Plot the chart

5 0 U C L = 3 9 . 2 4 0

C e n t r e l i n e 3 0 c = 2 4 . 4 2 0

1 0 L C L = 9 . 6 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5

Figure 5.12. The completed c-chart 4. Interpret the chart The above chart suggests that the upstream process which contributes to weaving faults in the fabric is not stable or in statistical control. There is evidence of several special causes - points outside the control limits. Improvement efforts should focus on identifying and eliminating the causes of instability in the weaving process.

Problem Solving and Process Improvement Tools and Techniques

Page 91

The u Chart
If we are monitoring occurrence data but the sampling interval/unit varies we can no longer use the c - chart, as this requires a constant sampling period/unit. For example, if a department store is monitoring the number of customer complaints per day and they have late night shopping on Friday and morning shopping on Saturday, the length of the day will not be the same for every day. A c-chart would not be used in such a situation. Where the data is occurrence data and the sampling unit/period is not constant a u-chart is used. The u-chart is to the c-chart what the p-chart is to the npchart.

u - Chart Example
The method for constructing a u-chart will now be illustrated with reference to the textiles example used for the c-chart. As part of the dying preparation process, the sections of fabric where significant weaving faults were found are cut out, and the remaining acceptable fabric sewn back into the roll. Rolls are also batched together or split into smaller rolls depending on the size of the customer order for a particular colour or pattern. This means that the rolls processed in the dying section will be of varying lengths depending on the customer order and the amount of flawed fabric that had to be cut out. Post-dying, the rolls are again inspected to identify any dying errors or faults. Sections with flaws or shade inconsistencies are cut out of the main roll and only fabric that meets the customers quality specifications are batched into the finished order, ready to be fabricated into finished items. eg. garments, bedlinen, curtains etc. Because our sampling interval or unit (the roll) is no longer constant, we can no longer use a c-chart to monitor the dying process. A u-chart is now the appropriate chart to use. Data collected for 25 consecutive customer orders is given in the following table.

Problem Solving and Process Improvement Tools and Techniques

Page 92

Table 5.15 Dying faults per customer order Order Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Total Roll length in Kilometres (n) 1.15 2.20 2.10 2.20 2.20 2.55 4.40 3.65 2.55 3.00 2.80 3.30 3.20 2.25 2.90 1.70 6.50 1.00 1.35 2.80 2.50 2.20 2.25 2.20 3.25 66.2 Number of dying faults (c) 15 18 23 22 18 15 25 19 13 33 29 27 32 21 26 17 36 11 14 26 25 24 20 15 18 542 Dying Faults per km (u=c/n) 13.0 8.2 11.0 10.0 8.2 5.9 5.7 5.2 5.1 11.0 10.4 8.2 10.0 9.3 9.0 10.0 5.5 11.0 10.4 9.3 10.0 10.9 8.9 6.8 5.5

Problem Solving and Process Improvement Tools and Techniques

Page 93

1. Calculate u for each subgroup of data collected. Calculate u for each unit or subgroup of data using the relationship: u = c/n

where n = size of the sampling unit or subgroup. For example, the first roll is 1.15 kilometres long and the number of faults is 15. Therefore n = 1.15 and c = 15. u is calculated as: u = c/n = 15/1.15 = 13 per kilometre.

2. Calculate the average nonconformities/unit (u). This can be calculated as follows: _ u =

c n
8.2

542 66.2

3. Calculate the Upper and Lower Control Limits Calculate the UCL and the LCL using the formula: _ _ UCL = u + 3 u ni _ _ LCL = u-3 u ni Note, the UCL and LCL will need to be calculated for each roll, to reflect the changing sample size. For example, for the first roll, the roll length is 1.15 kilometres. Therefore, UCL and LCL will be: UCL = 8.2 + 3 8.2 1.15 8.2 + 3 7.13 8.2 + 3 x 2.67 8.2 + 8.01 = 16.2

= = =

Problem Solving and Process Improvement Tools and Techniques

Page 94

LCL

8.2 - 3 8.2 1.15 8.2 - 8.01 0.19 = 0.2

= =

The completed UCL and LCL calculations for all the rolls (data points) are given in Table 5.16. Table 5.16 Dying faults per customer order, completed table Order Number Roll length in Kilometres (n) 1.15 2.20 2.10 2.20 2.20 2.55 4.40 3.65 2.55 3.00 2.80 3.30 3.20 2.25 2.90 1.70 6.50 1.00 1.35 2.80 2.50 2.20 2.25 2.20 3.25 66.2 Number of dying faults (c) 15 18 23 22 18 15 25 19 13 33 29 27 32 21 26 17 36 11 14 26 25 24 20 15 18 542 Dying Faults per km (u=c/n) 13.0 8.2 11.0 10.0 8.2 5.9 5.7 5.2 5.1 11.0 10.4 8.2 10.0 9.3 9.0 10.0 5.5 11.0 10.4 9.3 10.0 10.9 8.9 6.8 5.5 UCL LCL

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Total

16.2 14.0 14.1 14.0 14.0 13.6 12.3 12.7 13.6 13.2 13.3 12.9 13.0 13.9 13.3 14.8 11.6 16.8 15.6 13.3 13.6 14 13.9 14.0 13.0

0.2 2.4 2.3 2.4 2.4 2.8 4.1 3.7 2.8 3.2 3.1 3.5 3.4 2.5 3.2 1.6 4.8 - 0.4 0.8 3.1 2.8 2.4 2.5 2.4 3.4

Problem Solving and Process Improvement Tools and Techniques

Page 95

In the case of roll 18, the calculation for the LCL has given us a negative number. In practice we know that it is not possible to record a negative number of occurrences, therefore this would be recorded as a LCL of zero. 4. Draw the chart The completed chart is shown in Figure 5.13.
2 0

1 5

1 0

u = 8 . 2

0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5

Figure 5.13. The completed u chart 5. Interpret the chart. The chart suggests that the dying process is in control and centred around a mean of 8.2 occurrences of dying faults per kilometre of fabric dyed.

Problem Solving and Process Improvement Tools and Techniques

Page 96

Interpreting Control Charts

Several tests can be applied to Control Charts to assist with their interpretation. These are usually referred to as the "Shewhart tests". The foundation test is the presence of points beyond the control limits. The remaining tests are based on the fact that they represent events having the equivalent probability of occurring, in the absence of special causes, as a point beyond the control limits. The relevant tests are as follows: Points beyond the control limits Runs of points Trends Periodicity Hugging the centre-line Hugging the control limits

Points beyond the control limits. According to the laws of probability, a point has a 3 in 1,000 chance of falling outside the control limits purely by chance, for a stable process. The presence of any point falling outside the control limits, should initially be taken as indicating the presence of a special cause, ie. an indication that the process is unstable, and investigated as such. If the subsequent investigation fails to find a special cause, it can be assumed that what was seen was one of the 3 in 1,000. Runs of Points. The following runs of points are an indication of an unstable process. Eight consecutive points fall on one side of the centre-line; Two consecutive points fall more than 2 standard deviations from the centre-line on the same side of the centre-line; Two out of three consecutive points fall more than 2 standard deviations from the centre-line on the same side of the centre-line; Four out of five consecutive points fall more than 1 standard deviation from the centre-line on the same side of the centre-line.

Problem Solving and Process Improvement Tools and Techniques

Page 97

Trends. Any tendency of points to drift so as to give rise to a trend, is an indication that the process may be unstable. Periodicity. When points in a chart show a regular size and fall, the resulting pattern is referred to as periodicity. Periodicity is an indication that something in the process is changing in a regular pattern and therefore the process may be unstable. Hugging the centreline. If the data tends to hug the centreline, defined as 15 or more consecutive points less than 1 standard deviation from the centreline, an abnormality is indicated. This condition is usually as a result of data corruption, poor or inadequate sampling. Hugging the control limits. An abnormality is said to exist if 2 out of 3, 3 out of 7 or 4 out of 10 consecutive points fall between 2 and 3 standard deviations. Even if the points do not all fall on the same side of the centreline, this condition could be an indication that the process is unstable. The usual cause of this type of condition is unnecessary or overadjustment of the process. Shewharts original tests for control charts have been reproduced in Figure 5.14. In these diagrams, three zones have been defined above and below the mean, these are: Zone C - one standard deviation either side of the mean; Zone B - between one and two standard deviations away from the mean, one each side of the mean; Zone A - between two and three standard deviations out from the mean.

Problem Solving and Process Improvement Tools and Techniques

Page 98

Test 1. One point beyond zone A A B C C B A Test 3. Six points in a row steadily increasing or decreasing A B C C B A Test 5. Two out of three point in a row in zone A or beyond. A B C C B A Test 7. Fifteen point in a row in zone C (above & below the centreline) A B C C B A
Figure 5.14 The Shewhart tests

Test 2. Nine points in a row in Zone C or beyond A B C C B A Test 4. Fourteen points in a row alternating up and down A B C C B A Test 6. Four out of five points in a row in zone B or beyond. A B C C B A Test 8. Eight points in a row on both sides of centreline with none in zone C. A B C C B A

Problem Solving and Process Improvement Tools and Techniques

Page 99

Process Capability
Process capability is a measure of the ability of a process to meet or exceed the customer specifications for that process. Process capability is measured differently, depending on whether or not the process is centred on the mean of the customer specification. The term Cp is used where the process is centred, Cpk if it is not centred. For these measures to make any sense, the process must first be in statistical control and approximately normal.

Calculating Cp
When the process is centred on the mean of the customer requirement, process capability is measured by Cp . Where Cp is defined as follows: Cp = Specification width Process width

In practice, the customer requirements are defined in terms of a range from the lower specification limit to the upper specification limit. LSL and USL respectively. The process width is the 6 range defined between the mean plus three standard deviations to the mean minus three standard deviations. Process capability is therefore calculated by the formula: Cp = USL - LSL 6

Cp is interpreted as follows: A process with Cp = 1 exactly matches the customers specification limits; A process with Cp > 1 exceeds the customers specification; A process with Cp < 1 fails to meet the customers specification.

Problem Solving and Process Improvement Tools and Techniques

Page 100

Calculating Cpk
If a process is not centred on the mean of the customer requirement, process capability is measured by Cpk . Where Cpk is defined as follows: _ _ Cpk = Minimum of (USL - X , X - LSL ) 3 3 _ Where X is the mean of the process. The Cpk calculation overcomes the problem of the process not being centred by calculating the capability for each half of the process and then taking the minimum. Note: the upper one sided capability index is often referred to as Cpu and the lower one sided capability index as Cpl . ie. _ Cpu = USL - X 3 _ Cpl = X - LSL 3

Problem Solving and Process Improvement Tools and Techniques

Page 101

6. Extensions to the Seven Tools of Quality

This section presents a collection of tools and techniques that have all proved useful in support of quality based problem solving and continuous improvement activity, which are in addition to the seven basic tools of Quality. The list of tools that have been added as extensions to the basic quality management toolkit is ever growing as more of the many tools and techniques that were previously known about or have recently been developed are applied within a quality management framework, and in support of continuous improvement and problem solving activities. While by no means exhaustive, this section presents a collection of often used tools and techniques, which we have found be useful and practical extensions to the problem solving toolkit. These are: Interviews; Questionnaires; Dot Plots; Box Plots; Force Field Analysis.

Problem Solving and Process Improvement Tools and Techniques

Page 102

Interviews What are interviews?

Interviews are structured conversations focused on a specific topic. While usually more expensive than questionnaires, interviews allow longer responses and provide the interviewer with opportunities to pursue or ignore topics based on how the person being interviewed responds. Interviews also provide the interviewer with the opportunity to discover about issues that had not been previously thought of which are raised by the subject, and which may be of importance. Because interviews are conducted in person or over the telephone, they give the interviewer a much better "feel" for the subject's position on the topic.

When to use Interviews

Interviews can help teams get started on the problem solving process. Interviewing people in a department, work area or the whole organisation, will often reveal consistent themes about what people see as being major issues or problems. These can provide insights into areas that can be explored further using the problem solving process and that may offer greatest opportunities for improvement. Interviews can also be used to probe people's attitudes and opinions in depth, especially when there are relatively few people to survey, and/or when each person's opinions are very important.

How to use Interviews

1. Decide on the kind of information needed. Opinions, personal experiences, eye-witness accounts are good topics for interviews, as it may be difficult to obtain this information by other means. Avoid using interviews to obtain information available through other means. 2. Determine who you need to interview. Interviews are time consuming, therefore limit the number of people interviewed to the minimum required to find out what you want.

Problem Solving and Process Improvement Tools and Techniques

Page 103

3. Prepare questions in advance. Keep questions to the minimum required to obtain the desired information and try to keep interviews to under an hour; Balance your interview by mixing shorter, factual, easier, closed questions with open-ended, more thought provoking questions requiring longer answers; Test your questions on some people before conducting the actual interviews.

4. Prepare your interview subject ahead of time. Tell people in advance that you will be interviewing them and give them some idea about the sorts of things you will be asking, what the information will be used for, and how long the session is likely to take. Most people will want you to make a formal appointment with them. 5. Conduct the interview. When conducting the interview: Put your subject at ease by trying to keep the session as informal and conversational as possible; Review the purpose of the interview and the points raised in the preparation step above; Ask one question at a time and don't interrupt. Don't be afraid of silences; they can encourage people to volunteer more information. Remember, the purpose of interviewing is to listen to what your subject has to say - it should not be used as an excuse for you to tell people what you think; Take notes. If you use a tape recorder, be sure to ask for permission.

A note on using tape recorders It is usually more efficient to take notes, than to use a tape recorder. Especially when interviews are being conducted on a one-on-one basis as opposed to group interviews. Playing back to tapes to extract the key points which you would normally note during an interview can often take longer than doing the initial interviews. This significantly extends the time and effort required to extract the same information from your subjects. This can become particularly time consuming when there is a need to interview many people. Even with groups, many people will often feel threatened that if they something controversial, that the tape can be used to trace the comments back to them as individuals. Participants will often offer less input to an interview if it is being taped. Finally, not using a tape recorder will assist team members develop their skills as interviews and analysts. With time, team members will become very
Problem Solving and Process Improvement Tools and Techniques

Page 104

efficient at extracting information using the interview technique. Reliance on tape recorders can often work against team members fully developing these skills. Ask easier questions, closed questions and factual questions earlier in the interview; Ask more difficult open-ended questions later in the interview; Take opportunities during the interview to summarise and recap what your subject is saying; Allow for opportunities for the subject to raise and explore issues you may not have anticipated, this may give you additional insights; At the end of the interview, thank your subject for participating; After formally closing the interview, stay alert for any final remarks your subject may make. After the pressure is off, people often volunteer some of their most interesting observations and remarks.

6. As soon as possible after the interview, review your notes. Review your interview note, summarise the key points and any impressions you may have formed during the interview. 7. Assemble and analyse your responses. Depending on the purpose of the interview, you could tabulate responses for each question, sum up general reactions or identify recurring themes and issues.

Problem Solving and Process Improvement Tools and Techniques

Page 105

Questionnaires What is a Questionnaire ?

A questionnaire is a predetermined and structured proforma used to collect information about people's attitudes and opinions on specific issues or on specific dimensions of a problem or issue. Questionnaires may be completed in writing, verbally in person, or over the telephone.

When to use Questionnaires

Questionnaires can be used in any situations where you wish to collect attitudes and opinions about specific issues or problems, or where you wish to gauge peoples opinions and attitudes about some specific characteristics or dimensions of a problem or issues. Such cases lend themselves to pre-analysis and structuring. The team members can determine what issues? dimensions? characteristics? etc about which they seek opinions and attitudes. These can then be worked into a structured proforma questionnaire, on which subjects provided specific responses to specific questions.

How to use a Questionnaire

1. Determine what it is you need peoples opinions on. Determine the specific issues, problems, problem characteristics or dimensions you wish to gauge peoples opinions on. 2. Determine the who you need to survey. Determine which group or groups of people you need to survey. Do you need to survey customers, employees or only specific groups of employees eg. certain departments, certain shifts, certain sites. Do you need to compare the responses from different groups? or will different groups receive different questionnaires. 3. Prepare your questions. Use the following as guidelines to developing your questionnaire structure: Use open-ended questions when you're not certain of the nature of the response, when you need people's actual words or when you are looking for anecdotes or examples. Use close-ended questions when you want to focus responses on specific areas or issues. True-False and multiple choice answers are examples of this.

Problem Solving and Process Improvement Tools and Techniques

Page 106

Use a rating scale when you want to understand level of feeling. A commonly used format is to presented each question as a statement and the subject is asked to rate their level of agreement or disagreement with the statement. For example: I think the existing computer system response is too slow 1:strongly agree 2:agree 3: neutral 4: disagree 5: strongly agree

Use neutral language. Do not pre-empt a desired response or any bias through your wording. 4. Prepare your questionnaire. Some hints to assist your team in designing their questionnaires. Keep the number of questions to a minimum. When faced with lengthy questionnaires, many people ignore it completely, answer without much thought, or give more negative responses than they otherwise would because they are annoyed at having to answer such a long questionnaire - which is not indicative of how they really feel about the issues being probed by the questionnaire. If your questionnaire has a large number of questions, test it on a small group first. Responses many indicate many of your questions to be non-issues. These may be removed from your final questionnaire. Alternately, break up your questionnaire into several smaller questionnaires. Test your questionnaire first and revise any questions that are confusing or don't provide the required information. 5. Conduct the Survey When conducting the survey, using the questionnaire: If surveying a sample, ensure that it is a random sample. When surveying people from different groups, ensure the sample is representative of the numbers of people in the different groups. If possible, allow the responses to be anonymous for the individuals who complete the questionnaire - people will tend to be more honest with their answers; If you are surveying different groups of people, ensure you have allowed for some system of identifying which group a questionnaire came from with out breaching confidentiality.

Problem Solving and Process Improvement Tools and Techniques

Page 107

6. Analyse the results. Consolidate the responses to the categories established during the questionnaire designed stage. File the completed questionnaires is case this information is required for future reference.

Problem Solving and Process Improvement Tools and Techniques

Page 108

Dot Plots What are Dot Plots?

A dot plot is a pictorial way of displaying data. It is a plot of data points above a measurement scale with each point representing a single data point.

When to use Dot Plots

Use dot plots to get a quick idea as to the shape of the data you have collected. A glance at a dot plot will give you a good idea as to the shape, the range and spread, highlight any skewness, unusual or extreme points. This will often assist in deciding whether further analysis is worthwhile and what type of further analysis represents the logical next step. It allows easier interpretation than looking at a list of numbers.

How to Construct a Dot Plot

1. Determine what it is you need to investigate and collect your data. Example: Bank staff have complained to management that existing account withdrawal slips are too complicated. This results in frequent customer errors when filling out the form, errors which must be corrected by teller staff before the transaction can be processed. The time taken to make the correction is perceived as contributing to a decreasing level of customer service, forcing customers to wait while the corrections are made. Bank staff at one branch have decided to investigate this problem, by recording how many incorrectly filled withdrawal slips requiring teller staff correction are presented over a period of 30 working days. The results are as follows. 9 8 15 9 13 16 10 8 10 14 9 7 14 8 6 13 11 10 7 11 10 12 8 15 9 8 5 10 13 12

Problem Solving and Process Improvement Tools and Techniques

Page 109

2. Lay out a horizontal scale. Make sure the scale spans a large enough range to cover the range of the data. From examination of the above data, we can see that the largest number in the data is 16 and the smallest is 5. Therefore, a scale from 1 to 20 will more than cover the required range. This is illustrated in Figure 6.1.a.

1 2 3 4 5 1 0 1 1 1 2 1 3 6 7 8 9 1 4 1 5 1 6 1 7 1 8 1 9 2 0

Figure 6.1.a Layout of a Dot Plot Scale 3. Draw a dot over the number on your line corresponding to each value in your data for occurrence of that value. A partially completed dot plot for the above data is given in Figure 6.1.b.

1 2 3 4 5 6 8 9 1 0 1 1 1 2 1 3 7 1 4 1 5 1 6 1 7 1 8 1 9 2 0

Figure 6.1.b. A partially completed Dot plot Where a number is repeated several times in the data, the convention is to pile dots vertically on top of one another. For example, the number 10 occurs 5 times, the dot plot shows 5 dots over the value 10. Figure 6.1.c. shows the completed dot plot.

1 2 3 4 5 1 0 1 1 1 2 1 3 6 7 8 9 1 4 1 5 1 6 1 7 1 8 1 9 2 0

Figure 6.1.c. A completed Dot plot

Problem Solving and Process Improvement Tools and Techniques

Page 110

Box Plots What are Box Plots ?

A boxplot is a graphical way of displaying information about the spread and location of data. A boxplot focuses attention on certain features of the data without having to plot all the values. In particular, box plots are good at highlighting extreme points which are not typical of the rest of the sample and skewness in the data. Box plots provide a quick way of assessing data for which the team may be considering developing a control chart.

When to use Box Plots

Box plots can be used in the same types of situations where you would use a histogram or dot plot, to establish the location and spread of data. At the other end of the analytical scale, box plots may be used as an alternative to a control chart for one-off analytical exercises. One of the main advantages of box plots, are that we do not have to plot each individual data point. If we can come to some conclusion will a box plot, further and lengthier analysis will not be necessary. The other main advantage of a box plot, is that is can be used to determine whether or not there is a difference between two sets of data. This makes them extremely useful in situations where we seek to compare two or more sets of data. More so than histograms, dot plots and control charts - where apparent trends may be due to sampling errors or not statistically significant to indicate a real difference. Examples of this would include situations where we wish to establish whether data from to different sources are the same for the purpose of further analysis, or situations where we have implemented some improvements and we wish to establish whether the changes have made a difference or not by comparing before and after performance.

Problem Solving and Process Improvement Tools and Techniques

Page 111

How to construct a Box Plot

Unlike histograms and dot plots, box plots required us to have an understanding of some common terms and concepts associated with analysing and characterising data. In particular, we will need to know how to calculate: The median The mean The lower quartile The upper quartile

These terms and others associated with analysing and characterising data are presented and discussed in Appendix 1: "Understanding Variation and Data". Refer to this Appendix if you need a refresher on these concepts. 1. Define the problem to be investigated Example: A bank decides to investigate customer queue waiting times. 2. Collect data. The average waiting times for a random sample of customers joining a queue in front of a teller position are collected at five randomly selected times during the day, for a week. The average waiting time is calculated at each point. The data is given in Table 6.1. Table 6.1 Customer Queuing Times in Minutes (Averages of samples taken 5 times per day) Mon 6 4 5 12 5 Tues 1 10 2 7 3 Wed 2 6 9 5 8 Thurs 12 8 7 4 3 Fri 4 9 4 8 5

Sample 1 Sample 2 Sample 3 Sample 4 Sample 5

Problem Solving and Process Improvement Tools and Techniques

Page 112

3. Rearrange the data in ascending order To calculate the median, upper and lower quartiles, requires the data to be in order from the smallest value to the largest (ascending order) irrespective of what order the values occurred in when the measurements were taken. When you have only a few data points, it is relatively easy to re-order the data by examining it and rewriting it in the appropriate order. When there are many values, it is some what more difficult to do this by inspection. A tally chart is a useful tool to use to assist with the reordering of the data. To develop your tally chart, use the following procedure. i) Determine the spread of the data. From Table 6.1, the shortest waiting time is 1 minute (minimum data value), the longest is 12 minutes (maximum data value). ii) Set the spread of your tally chart. Set the minimum and maximum values for your tally chart as equal to the minimum and maximum values of your data. From Table 6.1, we will need a tally chart spanning the range from 1 (the minimum data value) to 12 (the maximum data value). iii) Set the class boundaries. The class boundaries need to be set to each of the discrete data intervals that are naturally occurring in the data. iv) Fill in the Tally Chart from your data. Analyse the data given in Table 6.1 with the tally chart. The completed Tally chart for the data given in Table 6.1 is shown as Table 6.2. Table 6.2 Tally chart for data in Table 6.1 Waiting Time (Minutes) 1 2 3 4 5 6 7 8 9 10 12 Number of Occurrences Number of Occurrences (Tally) I 1 II 2 II 2 III 3 I I I I5 II 2 II 2 III 3 II 2 I 1 II 2

Problem Solving and Process Improvement Tools and Techniques

Page 113

Total 25 The table tells us how many times each measured value occurred in the collected data. Using this information we can rewrite the data collected in ascending order, as is shown in Table 6.3. This table shows that we have 25 data points, the lowest being 1 and the highest being 12. With reference to the tally chart (Table 6.2), we note that the value 5 minutes was recorded on 5 occasions during the week. This is shown in Table 6.3 as a block of 5 consecutive 5s. Similarly, the tally chart shows that the value 8 minutes was observed on three occasions. This appears on Table 6.3 as a block of three 8s. Table 6.3. Table 6.1 data, reordered in ascending order Position 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Measured Value 1 2 2 3 3 4 4 4 5 5 5 5 5 6 6 7 7 8 8 8 9 9 10 12 12

Lower quartile (middle of the lower half)

Median (the middle value)

Upper quartile (middle of upper half)

Problem Solving and Process Improvement Tools and Techniques

Page 114

4. Calculate the value required to draw the Box Plot. The key values we need to calculate in order to draw a box plot are the: Median; Upper Quartile; Lower Quartile; Interquartile Range; Mean.

The mean, upper and lower quartiles can be determined from Table 6.3 by observations. i) Determine the Median The median is the middle value in the data. We have 25 data values, the middle value occupies the 13th position. ie. it has 12 values either side of it. The 13th data point, the median value is 5 minutes. ii) Determine the Lower Quartile The lower quartile is the middle value of the lower half of the data. The usual convention is to include the overall median as the upper limit of the bottom half of the data. The middle value of the lower half of the data is the 7th value. ie. It has 6 data points below it, 1 to 6 and 6 data points above it, 8 to 13. The data value corresponding to the 7th point is 4 minutes. Therefore, the lower quartile is 4 minutes. iii) Determine the Upper Quartile The upper quartile is the middle value of the upper half of the data. The usual convention is to include the overall median as the lower limit of the upper half of the data. From Table 6.3. we can see that the middle value of the top half of the data is the 19th data point which corresponds to the value of 8 minutes. Therefore, the upper quartile is 8 minutes.

Problem Solving and Process Improvement Tools and Techniques

Page 115

iv) Calculate the Interquartile Range The Interquartile range (IQR) is the difference between the Upper Quartile and the Lower Quartile. That is: IQR IQR = = Upper Quartile 8 Minutes Lower Quartile = 4 Minutes.

4 Minutes

Interquartile Range = 4 minutes. v) Calculate the mean. Add all the data points ie. 1+2+2+3+3+4+ ..... +12 = Then divide by the total number of points (25) ie. 150/25 The mean = 6 5. Draw the Box Plot Draw the Box Plot using the following steps and with reference to the information previously calculated. i) Lay out a horizontal axis and mark off a measurement scale in the units you used to measure your data. Ensure the line is drawn wider than the range of your data. This is shown in Figure 6.2.a. 150 = 6

0 1 2 3 4 5 6 7 8 9 1 0 1 2 1 3 1 1
Figure 6.2.a. Layout a horizontal scale ii) Mark off the Lower and Upper Quartiles with vertical lines of the same length and connect them up to form a box shape. See Figure 6.2.b.

0 1 2 3 4 5 6 7 8 9 1 0 1 2 1 3 1 1
L o w e r Q u a r t i l e U p p e r Q u a r t i l e

Figure 6.2.b. Drawing the Upper and Lower Quartiles

Problem Solving and Process Improvement Tools and Techniques

Page 116

iii) Calculate and draw the Lower Inner Fence Measure to a length of one-and-a-half interquartile ranges below the lower quartile. Mark this with a dashed vertical line. This is called the lower inner fence. In our example, 1.5 X IQR = 6. Therefore the lower inner fence is calculated as the lower quartile minus six ie 4 - 6 = -2. Note, this is a calculated value, in practice that we cannot have a waiting time of less than zero minutes. iv) Calculate and draw the Upper Inner Fence Measure to a length of one-and-a-half interquartile ranges above the upper quartile. Mark this with a dashed vertical line. This is called the upper inner fence. In our example, the upper inner fence is calculated as the upper quartile plus six. ie. 8 + 6 = 14. v) Mark in the lowest & highest data points inside the fences Mark in the lowest data point that lies inside the lower inner fence with a dot. Draw the dot in-line with the centre-line of the box and connect it to the box by a straight line. Mark in the highest data point that lies inside the upper inner fence with a dot and connect it to the box by a straight line. See Figure 6.2.c. The lines connecting the box to the dots are called whiskers.
U p p e r I n n e r F e n c e

L o w e r I n n e r F e n c e

W h i s k e r s

W h i s k e r s

1 3 1 4 0 1 2 3 4 5 6 7 8 9 1 0 1 2 2 1 1 1
L o w e r Q u a r t i l e U p p e r Q u a r t i l e

Figure 6.2.c. Box plot with Inner Fences and Whiskers vi) Mark in the median and the mean Mark the median in the box with cross (+) and the mean with a short straight. This is illustrated by Figure 6.2.d which shows the completed box plot.

Problem Solving and Process Improvement Tools and Techniques

Page 117

L o w e r I n n e r F e n c e

U p p e r I n n e r F e n c e

W h i s k e r s

W h i s k e r s

1 3 1 4 0 1 2 3 4 5 6 7 8 9 1 0 1 2 2 1 1 1
L o w e r Q u a r t i l e M e a n M e d i a n U p p e r Q u a r t i l e

Figure 6.2.d. The completed Box Plot vii) Mark in any outliers. If data point occur which lie outside the inner fences, mark these with an asterisk *. Suppose one of our measurements was 16 minutes. This would be plotted as shown in Figure 6.2.e. Data points lying outside the inner fences are called outliers.
U p p e r I n n e r F e n c e O u t l i e r W h i s k e r s W h i s k e r s

L o w e r I n n e r F e n c e

1 3 1 4 1 6 0 1 2 3 4 5 6 7 8 9 1 0 1 2 1 5 2 1 1 1
L o w e r Q u a r t i l e M e a n M e d i a n U p p e r Q u a r t i l e

Figure 6.2.e. Box Plot showing an Outlier

Problem Solving and Process Improvement Tools and Techniques

Page 118

Interpreting a Box Plot

The following points provide some guidance to interpreting a box plot. The inner fences provide a measure of the likely spread of other samples taken from the same population. If the data is normally distributed, 99.3 % of the data will lie within the inner fences. 99.9% of the data will lie within three interquartile ranges of the quartiles. This measure is referred to as the outer fences. If the data is normally distributed, the median will be in the centre of the box; the mean will be the same as the median; the whiskers will be of even length;

The degree to which the median is off-centre gives a good indication as to the degree to which the data is skewed on non-normal. Similarly, whiskers of uneven length are also an indication that data is skewed. Points outside the inner fences or "outliers" are unusual and should be investigated to determine their cause. They indicate that some change, aberration or unusual event has caused this particular point to be very different from the usual performance of the process we are measuring. To fairly assess the performance of the process, outliers should be excluded as they would otherwise excessively skew the data. In summary, a box plot will show at a glance: The range of the data The medians and quartiles The interquartile range Whether or not data is skewed (a median that is off-centre or whiskers of uneven length) Outliers Whether there is a difference between groups of data.

Using Box Plots to compare groups of data

Box plots can be used to test whether there is a real difference between two groups of data, or whether apparent differences are due to random variations in the process.
Problem Solving and Process Improvement Tools and Techniques

Page 119

For example, a company monitoring customer satisfaction can use a box plot to determine whether a recent improvements in satisfaction levels is due to a genuine improvement in customer satisfaction or simply due to the fact that the group sampled was biased towards "happy" customers and as such did not represent a "fair" sample. ie. this is usually referred to a sampling error. In the bank queuing time example previously considered, suppose the bank staff implemented some changes aimed at reducing waiting time. A second group of data could be collected after the changes. By drawing a box plot for the second group (after) data on the same graph as the first box plot we, can assess whether or not the changes did actually reduce waiting time. Figure 6.2.f. shows the second group of data collected after the changes, plotted with the original box plot and illustrates graphically the value of box plots for comparing groups of data.

N o o v e r l a p i n d i c a t e s a " r e a l " d i f f e r e n c e b e t w e e n t w o s e t s o f d a t a

A f t e r t h e c h a n g e s

O r i g i n a l D a t a

1 3 1 4 1 6 0 1 2 3 4 5 6 7 8 9 1 0 1 2 1 5 2 1 1 1

Figure 6.2.f. Using Box Plots to compare two groups of data If the boxes do not overlap, as is the case in Figure 6.2.f, then we can say that there is a real difference between the two groups. If the boxes overlap to any degree, then any apparent change in the data is due to sampling errors and not due to any fundamental changes or differences in the processes which the data is measuring.

Problem Solving and Process Improvement Tools and Techniques

Page 120

Force Field Analysis What is Force Field Analysis?

Force field analysis is a graphic technique that allows a team to display the forces acting for and against the successful implementation of any proposed change or countermeasure. Also known as a barriers and aids analysis.

When to use a Force Field Analysis?

Force field analysis is most useful when formulating an implementation strategy to ensure the team have comprehensively addressed all of the key issues. Force field analysis is particularly good at identifying and highlighting both problem areas and likely areas of resistance as well as key points of leverage which can be exploited by the team to overcome resistance and other implementation problems.

How to do a Force Field Analysis

1. Decide on and describe the goal or change to be analysed. Decide on and describe the goal or change to be analysed. Write this on a sheet of paper or white board and draw a box around it. Draw a line across the page starting at the box to act as the dividing line between positive and negative forces. This is illustrated in Figure 6.3.a.

Improve delivery performance

Figure 6.3.a. Starting layout for force field analysis. 2. Brainstorm and list all of the positive factors. Brainstorm and list all of the factors which will support the achievement of the goal or implementation of the change. Record these on the force field analysis diagram with arrow pointing toward the horizontal line. This is illustrated in Figure 6.3.b.
Problem Solving and Process Improvement Tools and Techniques

Page 121

Improve delivery performance

Customer demands

Reduction in WIP inventories

Less hassles with customer complaints

Reduce conflict between departments

Reduce transport costs

Figure 6.3.b. Record the positive forces supporting the change 3. Brainstorm and list all of the negative factors Brainstorm and list all of the factors which will oppose the achievement of the goal or implementation of the change. Record these on the force field analysis diagram with arrow pointing toward the horizontal line on the opposite side from the positive forces. This is illustrated in Figure 6.3.c.
Resistance changes to Resistance to credit policy use of new by accounts procedures Cost of new computer hardware Getting agreement to changes from the bank

Improve delivery performance

Customer demands

Reduction in WIP inventories

Less hassles with customer complaints

Reduce conflict between departments

Reduce transport costs

6.3.c. Force field diagram with negative factors added. 4. Highlight the most significant forces Highlight the forces the team believe will be the most significant by drawing a larger arrow and stonger line from these. See Figure 6.3.d.

Problem Solving and Process Improvement Tools and Techniques

Page 122

Resistance changes to Resistance to credit policy use of new by accounts procedures

Cost of new computer hardware

Getting agreement to changes from the bank

Improve delivery performance

Customer demands

Reduction in WIP inventories

Less hassles with customer complaints

Reduce conflict between departments

Reduce transport costs

6.3.d. The completed force field analysis. 5. Assess the results Assess the results of the force field analysis and modify your implementation strategy to take into account the new information.

Problem Solving and Process Improvement Tools and Techniques

Page 123

7. The Seven Quality Management and Planning Tools

The seven quality management and planning tools were identified during the early 1970s by the Japanese Society of Quality Control and Technique Development, as part of a deliberate initiative aimed at identifying and evaluating management control techniques which had been proven to be the most effective. They identified seven new management techniques. These are: The affinity diagram; Interrelationship diagram; Systematic diagram; Matrix Diagram/Decision Matrix; Prioritisation Matrix; Process decision program chart; Activity network diagram;

In contrast to the seven basic quality tools, which are predominantly quantitative in nature, based on the use of hard data and applied at the shopfloor level, the seven quality management and planning tools are aimed at problems that are predominantly qualitative. They are mainly used by the middle and upper management levels to help identify problems, recommend and establish plans for corrective action. The relationship between the different tools when they are used as part of an integrated approach to address some management issue or problem is illustrated in Figure 7.1.

Problem Solving and Process Improvement Tools and Techniques

Page 124

R e l i e s o n c r e a t i v i t y

R e l i e s o n l o g i c

S t a g e 1

A f f i n i t y D i a g r a m

R e l a t i o n s D i a g r a m

S t a g e 2 D e c i s i o n M a t r i x P r i o r i t i s a t i o n M a t r i x M a t r i x D i a g r a m S y s t e m a t i c D i a g r a m

S t a g e 3

A r e t a s k s k n o w n ?

N e t w o r k D i a g r a m

P r o c e s s D e c i s i o n P r o g r a m C h a r t

Figure 7.3. The Seven Quality Management & Planning Tools

Figure 7.1 illustrates the various tools used during three major stages, as follows: Stage 1: to identify and clarify the nature of the problem, the key issues and dimensions; Stages 2: to identify and clarify the relationships and interactions between issues and relationship to the objective; Stage 3: to develop time phased plans for implementation, addressing risk and developing contingencies. This section provides a description of the usage of each of these tools and techniques.

Problem Solving and Process Improvement Tools and Techniques

Page 125

The Affinity Diagram What is an Affinity Diagram?

The affinity diagram is a means of collecting a large amount of verbal data and organising it into natural clusters. The clusters will ideally represent the intrinsic structure of the problem being investigated. The process of creating an affinity diagram promotes the generation, organisation and consolidation of information about a problem or other complex issue. The process of developing affinity diagrams is also known as the KJ method (its Japanese name), after its developer Mr Jiro Kawakita, a Japanese anthropologist.

When do you use an Affinity Diagram?

The strength of the affinity diagram is in the way it can help with unravelling and understanding complex problems and issues. It is useful for collecting and consolidating team members thoughts on unknown or unexplored areas. In particular, the affinity diagram is best used when dealing with problems and issues that: are complex in nature and difficult to get your mind around; are disorganised or chaotic; have resisted attempts to resolve them using other approaches eg. the seven basic tools and problem solving process; require a large number of ideas and inputs, all of which must be organised; require a degree of creativity if they are going to be solved; required team efforts.

How do you use an Affinity Diagram?

Identify the problem or issue to be addressed. Consider an organisation which has had ongoing problems meeting delivery targets on customer orders . Significant internal conflict has arisen as managers and staff in the various departments involved in the order fulfilment process try to assign blame for the delays to each others department.

Problem Solving and Process Improvement Tools and Techniques

Page 126

The organisation has decided to examine the order fulfilment process to identify and resolve the problems that affect the organisations delivery performance. Nominate a team leader or facilitator to lead the session. Developing an affinity diagram requires strong facilitation. The team should nominate a team member or if necessary, invite some one external to the team who has a good command of the affinity diagram development process to lead the session(s). Develop a problem or issues statement. Prepare a formal problem or issue statement, to focus the attention of the team - this is often best done by stating the problem or issue in the form of a question eg. how do we accelerate our product development cycle? or what are the issues associated with implementing this new technology? In our case, What are the problems that impact on customer delivery performance. Individual and Team Brainstorming. The brainstorming phase can be progressed as a group brainstorming session, as individual brainstorming or as a combination of both. The key part is that as each idea is thought of or verbalised, it is written on a small piece of paper or card. (Post-it notes are particularly good for this). One approach is to have each person brainstorm individually and list their ideas on a sheet of paper. When this is done, the session leader has every one offer one idea . This idea is written on a piece of paper or card (one idea per card) and the card is placed on the table. This process continues until everyone has exhausted their ideas. The same rules that apply to brainstorming are applied. Any ideas that are thought of as this process progresses are written down on cards and added to the pile.

Problem Solving and Process Improvement Tools and Techniques

Page 127

N o I n s t a l l a t i o n D e l i v e r y a d d r e s s N o I n s t a l l a t i o n N o I n s t a l l a t i o n O r d e r h e l d u p i n M a n u a l i n B o x w r o n g N o I n s t a l l a t i o n M a n u a l i n B o x N o t r a n s p o r t O r d e r h e l d u p i n M a n u a l i n B o x C r e d i t c h e c k t a k e s i n t e r n a l m a i l M a n u a l i n B o x N o t r a n s p o r t o r g a n i s e d C r e d i t c h e c k t a k e s i n t e r n a l m a i l N o t l e a r w h a t t o o l o n g N o I n s t a l l a t i o n D e l i v e r y a d d r e s s o r g a n i s e d N o t c l e a r w h a t N o I n s t a l l a t i o n t o o l o n g c u s t m e r w a n t s O f f i c i a l o r d e r N o I n s t a l l a t i o n O r d e r h e l d u p i n M a n u a l i n B o x w r o n g N o I n s t a l l a t i o n c u s t o m e r w a n t s M a n u a l i n B o x B i c k e r i n g b e t w e e n N o t r a n s p o r t O r d e r h e l d u p i n i n c o r e c t M a n u a l i n B o x C r e d i t c h e c k t a k e s i n t e r n a l m a i l M a n u a l i n B o x N o t r a n s p o r t d e p a r t m e n t s o r g a n i s e d C r e d i t c h e c k t a k e s i n t e r n a l m a i l t o o l o n g T e c h n i c i a n s b u s y o r g a n i s e d N o t c l e a r w h a t t o o l o n g B i c k e r i n g b e t w e e n o n o t h e r w o r k c u s t o m e r w a n t s B i c k e r i n g b e t w e e n d e p a r t m e n t s d e p a r t m e n t s

Consolidate cards into like piles. Spread all the cards out on a table. Every individual in the group then participates in arranging the cards into similar or related groups ie. cards that have an affinity to each other. Some guidelines to conducting this exercise include: While grouping the cards into like piles, there should be no talking between members; Cards can be moved between piles by different team members as many times as required until every one is happy with the arrangement; Cards that do not seem to fit in any one pile can be grouped into a miscellaneous pile. The exercise is finished when team members are no longer moving cards between piles.

Problem Solving and Process Improvement Tools and Techniques

Page 128

O f f i c i a lo r d e r i n c o r r e c t S t o c k c o d e i n c i r r e c t , c a n n o t i n p u t N e w c u s t o m e r , n o d e t a i l s o n c o m p u t e r E t c . , E t c . ,

T e c h n i c i a n s b u s y o n o t h e r w o r k N o t e c h n i c i a n a l l o c a t e d T e c h n i c i a n n o t a v a i l a b l e E t c . , E t c . , I t e m d a m a g e d W r o n g m a n u a li n b o x I t e m d o e s n ' t w o r k I n c o r r e c t i t e m d e l i v e r e d E t c . , E t c . , N o i n s t a l l a t i o n m a n u a li n b o x P a r t s m i s s i n g f r o m i n s t a l l a t i o n k i t E t c . , E t c . , N o t r a n s p o r t o r g a n i s e d W a i t i n g o n b a n k a p p r o v a l N o s u i t a b l e t r u c k a v a i l a b l e W a i t f o r a v a i l a b l e s p a c e o n n e x t t r u c k E t c . , E t c . , B i c k e r i n g b e t w e e n d e p a r t m e n t s
N o b o d y h e l p s o u t w h e n t h e r e i s a p r o b l e m

O r d e r i n i n t e r n a l m a i l s y s t e m

N o t t y p e d i n t o c o m p u t e r O f f i c i a lo r d e r n o t m a d e u p E t c . , E t c . ,

E q u i p m e n t s i t t i n g o n d e s p a t c h d o c k
l o a d i n g d o c k s j a m m e d w i t h o t h e r p r o d u c t

I t e m n o t i n s t o c k , n e e d t o p u r c h a s e

T u r n a r o u n d i n a n n o t f i n d c r e d i t I t e m w a i t i n g t o b e d e s p a t c h t o o l o n g C d e t a i l s p i c k e d E t c . , E t c . , B a n k e r n o t B a c k l o g i n a v a i l a b l e w a r e h o u s e t o o h i g h E t c . , E t c . , E t c . , E t c . ,

N o c o m m u n i c a t i o n o n p r i o r i t i e s E t c . , E t c . ,

Figure 7.4. Cards segregated into piles with similar or related cards Allocate a title to each pile. Examine the cards in each pile. Based on the theme of the cards in each pile, allocate titles to the piles which are descriptive of the theme of ideas in the pile. This can be progressed through group discussion. The titles for each pile are then written on a card which is placed at the top of the pile as a header. Note: the most useful titles are usually those expressed in the form of a short phrase describing the theme of the pile rather than a single word title.

Problem Solving and Process Improvement Tools and Techniques

Page 129

Orde r in inte rnal mail syste m

Official order incorrect Stock code incirrect, cannot input N ew customer, no details on computer Etc., Etc.,

Technicians busy on other work N o technician allocated Technician not available Etc., Etc.,

I tem damaged I tem doesn' t work I ncorrect item delivered Etc., Etc.,

Wrong manual in box N o installation manual in box Parts missing from installation kit Etc., Etc.,

N ot typed into computer Official order not made up Etc., Etc.,

I tem not in stock, need to purchase I tem waiting to be picked Backlog in warehouse too high Etc., Etc.,

Equipment sitting on despatch dock

loading docks j amme d with othe r product

Waiting on bank approval Cannot find credit details Banker not available Etc., Etc.,

N o transport organised N o suitable truck available Wait for available space on next truck Etc., Etc.,

Bickering between departments

N o body he lps out whe n the re is a proble m

Turnaround in despatch too long Etc., Etc.,

N o communication on priorities Etc., Etc.,

Figure 7.5. Piles with title cards Review the miscellaneous pile. Review the miscellaneous pile to see if any of the cards can be reallocated to the named piles. Draw the affinity diagram. Use adhesive tape to position the cards, in the groupings that emerged on a large flip chart or pieces of butchers paper which can be stuck together, to shown the whole diagram. The header card should be at the top of each section.

Problem Solving and Process Improvement Tools and Techniques

Page 130

W h a t a r e t h e p r o b l e m s t h a t i m p a c t o n c u s t o m e r d e l i v e r y p e r f o r m a n c e

P i c k i n g D e l a y s

D e l a y s w i t h C r e d i t C h e c k

T r a n s p o r t s c h e d u l i n g & c o o r d i n a t i o n

O r d e r h e l d u p i n t h e " s y s t e m "

S c h e d u l i n g o f t e c h n i c i a n s

C a n n o t i n s t a l l , i t e m s m i s s i n g

I t e m n o t i n s t o c k , n e e d t o p u r c h a s e I t e m w a i t i n g t o b e p i c k e d B a c k l o g i n w a r e h o u s e t o o h i g h E t c . , E t c . ,

W a i t i n g o n b a n k a p p r o v a l C a n n o t f i n d c r e d i t d e t a i l s B a n k e r n o t a v a i l a b l e

N o t r a n s p o r t o r g a n i s e d

O r d e r i n i n t e r n a l m a i l s y s t e m N o t t y p e d i n t o c o m p u t e r

T e c h n i c i a n s b u s y o n o t h e r w o r k N o t e c h n i c i a n a l l o c a t e d

W r o n g m a n u a l i n b o x

N o s u i t a b l e t r u c k a v a i l a b l e W a i t f o r a v a i l a b l e s p a c e o n n e x t t r u c k E t c . , E t c . ,

N o i n s t a l l a t i o n m a n u a l i n b o x P a r t s m i s s i n g f r o m i n s t a l l a t i o n k i t E t c . , E t c . ,

O f f i c i a l o r d e r n o t m a d e u p E t c . , E t c . ,

T e c h n i c i a n n o t a v a i l a b l e E t c . , E t c . ,

E t c . , E t c . ,

O r d e r h e l d u p i n D e s p a t c h

L a c k o f C o o p e r a t i o n b e t w e e n D e p t s .

N o t e n o u g h i n f o r m a t i o n c a p t u r e d o n o r d e r

I t e m n o t f u n c t i o n a l

E q u i p m e n t s i t t i n g o n d e s p a t c h d o c k l o a d i n g d o c k s j a m m e d w i t h o t h e r p r o d u c t T u r n a r o u n d i n d e s p a t c h t o o l o n g E t c . , E t c . ,

B i c k e r i n g b e t w e e n d e p a r t m e n t s N o b o d y h e l p s o u t w h e n t h e r e i s a p r o b l e m N o c o m m u n i c a t i o n o n p r i o r i t i e s E t c . , E t c . ,

O f f i c i a l o r d e r i n c o r r e c t

I t e m d a m a g e d

S t o c k c o d e i n c i r r e c t , c a n n o t i n p u t N e w c u s t o m e r , n o d e t a i l s o n c o m p u t e r E t c . , E t c . ,

I t e m d o e s n ' t w o r k

I n c o r r e c t i t e m d e l i v e r e d

E t c . , E t c . ,

Figure 7.6. The completed diagram Evaluate, assess and discuss the affinity diagram. The team members now assess and discuss the diagram, the categories or titles in which the cards grouped, the individual cards in each pile, in order to gain more insight into the problem or issue. The grouping categories should flag the major dimensions or components of the problem or issues - understanding these will add greater structure and focus to activities aimed at addressing these issues.

What next?
The affinity diagram can greatly enhance a teams understanding of a problem or issue, by flagging the major elements. However, it may still not be obvious what to do next. This is where the interrelationship diagram can be used to build on the outputs of the affinity diagram, by identify which element may be more important or exerts undue influence on the problem. These areas, once identified should provide the initial point of focus for improvement efforts.

Problem Solving and Process Improvement Tools and Techniques

Page 131

Relations Diagram What is a Relations Diagram?

The relations diagram is a graphical method for clarifying and representing complex cause and effect relationships. It illustrates the logic that exists between the many factors or elements of a problem or issue. This technique may also be used for displaying complex objectives means-toends relationships.

When do you use a Relations Diagram?

The relations diagram is used to determine the root causes and root effects of a problem. Root causes are those factors or elements of a problem that generate the majority of the symptoms. Root effects are the major symptoms. The relations diagram is particularly useful for determining the root causes of a problem or issues in cases where quantified data is not available, or where there is a large number of interrelated issues to be defined and understood. Where there are a large number of factors to be understood, it is often useful to use the affinity diagram as a method for identifying the key factors or issues, which are then used as an input to the relations diagram.

How do you use a Relations Diagram?

Define the problem or issue to be addressed. Agree a problem or issue statement. Prepare a short problem or issue statement to focus the attention of the team - this is often best done by stating the problem or issue in the form of a question eg. how do we accelerate our product development cycle? or what are the issues associated with implementing this new technology? In our case, What are the problems that impact on customer delivery performance. Layout the major factors/elements. Write the problem statement in the centre of a large sheet of paper or chart. Draw a circle around the problem statement.

Problem Solving and Process Improvement Tools and Techniques

Page 132

Lay out the key factors or elements of the problem, previously identified using the affinity diagram evenly around the problem statement. This is illustrated in Figure 7.5.
C a n n o t i n s t a l l , i t e m s m i s s i n g D e l a y s w i t h C r e d i t C h e c k O r d e r h e l d u p i n D e s p a t c h

N o t e n o u g h i n f o r m a t i o n c a p t u r e d o n o r d e r

I t e m n o t f u n c t i o n a l

W h a t a r e t h e p r o b l e m s t h a t i m p a c t o n c u s t o m e r d e l i v e r y p e r f o r m a n c e
T r a n s p o r t s c h e d u l i n g & c o o r d i n a t i o n L a c k o f C o o p e r a t i o n b e t w e e n D e p t s .

O r d e r h e l d u p i n t h e " s y s t e m "
P i c k i n g D e l a y s

S c h e d u l i n g o f t e c h n i c i a n s

Figure 7.7. Starting layout for a relations diagram Analyse the relationships Consider each factor or element in the context of the other factors/elements. Assess whether or not a cause or effect relationship exists between the factors. ie. does this factor/element influence any other? Is this factor/element the result of or affected by any other? If a relationship exists between two factors, draw a line between them. Draw an arrowhead on the end of the line in the direction of the effect relationship. For example, in Figure 7.6, the factor Credit Check Delays was assessed as being caused by the factors Not enough information captured on order. The line is draw with the arrow pointing to Credit Check Delays.

Problem Solving and Process Improvement Tools and Techniques

Page 133

C a n n o t i n s t a l l , i t e m s m i s s i n g D e l a y s w i t h C r e d i t C h e c k O r d e r h e l d u p i n D e s p a t c h

N o t e n o u g h i n f o r m a t i o n c a p t u r e d o n o r d e r

I t e m n o t f u n c t i o n a l

W h a t a r e t h e p r o b l e m s t h a t i m p a c t o n c u s t o m e r d e l i v e r y p e r f o r m a n c e
T r a n s p o r t s c h e d u l i n g & c o o r d i n a t i o n L a c k o f C o o p e r a t i o n b e t w e e n D e p t s .

O r d e r h e l d u p i n t h e " s y s t e m "
P i c k i n g D e l a y s

S c h e d u l i n g o f t e c h n i c i a n s

Figure 7.8. Arrow illustrating cause-and-effect relationship The team continues to analyse the relationships until all factors have been compare to all others. The result of this analysis is shown in Figure 7.7.
C a n n o t i n s t a l l , i t e m s m i s s i n g D e l a y s w i t h C r e d i t C h e c k O r d e r h e l d u p i n D e s p a t c h

N o t e n o u g h i n f o r m a t i o n c a p t u r e d o n o r d e r

I t e m n o t f u n c t i o n a l

W h a t a r e t h e p r o b l e m s t h a t i m p a c t o n c u s t o m e r d e l i v e r y p e r f o r m a n c e
T r a n s p o r t s c h e d u l i n g & c o o r d i n a t i o n L a c k o f C o o p e r a t i o n b e t w e e n D e p t s .

O r d e r h e l d u p i n t h e " s y s t e m "
P i c k i n g D e l a y s

S c h e d u l i n g o f t e c h n i c i a n s

Figure 7.9. A completed analysis of relations

Problem Solving and Process Improvement Tools and Techniques

Page 134

Count the numbers of arrows in each direction. For each box defining a factor or element. Count the number of arrows going in and the number of lines going out. Label each box with these numbers in the format arrows in/arrows out. This is illustrated in Figure 7.8.
2 / 1 1 / 2
D e l a y s w i t h C r e d i t C h e c k C a n n o t i n s t a l l , i t e m s m i s s i n g

a r r o w s i n / a r r o w s o u t 2 / 1
O r d e r h e l d u p i n D e s p a t c h

0 / 6
N o t e n o u g h i n f o r m a t i o n c a p t u r e d o n o r d e r

1 / 1
I t e m n o t f u n c t i o n a l

4 / 2
T r a n s p o r t s c h e d u l i n g & c o o r d i n a t i o n

W h a t a r e t h e p r o b l e m s t h a t i m p a c t o n c u s t o m e r d e l i v e r y p e r f o r m a n c e

4 / 1
L a c k o f C o o p e r a t i o n b e t w e e n D e p t s .

5 / 2
O r d e r h e l d u p i n t h e " s y s t e m "
P i c k i n g D e l a y s

2 / 3
S c h e d u l i n g o f t e c h n i c i a n s

3 / 5

Figure 7.10. Relations diagram with arrows-in/arrow-out annotation Identify root causes and root effects The root causes and root effects are identified by assessing the arrowsin/arrows-out count. The root causes will be the factors/elements with the most arrows going out; The root effects, those with the most arrow going in; The primary root cause will be the root cause with the highest number of arrows going out; The primary root effect will be the one with the highest number of arrows going in.

In the same manner as the Pareto chart attempts to identify root causes for the team to focus their attention on, the relations diagram performs the same function in situations where the quantitative data may not be available to construct a Pareto chart.

Problem Solving and Process Improvement Tools and Techniques

Page 135

However, the team will still need to reflect on their own experience and use common sense. If there are factors that are known by the team to be significant causes, these will need to be considered, even if they are not specifically highlighted by the relations diagram. A completed relations diagram with root causes and root effects highlighted is shown in Figure 7.9.
2 / 1 1 / 2
D e l a y s w i t h C r e d i t C h e c k C a n n o t i n s t a l l , i t e m s m i s s i n g

a r r o w s i n / a r r o w s o u t 2 / 1
O r d e r h e l d u p i n D e s p a t c h

0 / 6
N o t e n o u g h i n f o r m a t i o n c a p t u r e d o n o r d e r

1 / 1
I t e m n o t f u n c t i o n a l

4 / 2
T r a n s p o r t s c h e d u l i n g & c o o r d i n a t i o n

W h a t a r e t h e p r o b l e m s t h a t i m p a c t o n c u s t o m e r d e l i v e r y p e r f o r m a n c e

4 / 1
L a c k o f C o o p e r a t i o n b e t w e e n D e p t s .

5 / 2
O r d e r h e l d u p i n t h e " s y s t e m "
P i c k i n g D e l a y s

2 / 3
S c h e d u l i n g o f t e c h n i c i a n s

R o o t C a u s e 3 / 5 R o o t E f f e c t

Figure 7.11. Root Cause and Root Effect Identified Assess the final diagram. Redraw the final diagram neatly and assess the results. Do they make sense? Were there any surprises?

From our diagram we can see that: The primary root cause has been identified as Not enough information captured on order. Other key causes include Picking delays and Scheduling of Technicians. The primary root effect is Orders held up in the system. Other key effects are Transport scheduling and coordination and Lack of cooperation between departments.
Problem Solving and Process Improvement Tools and Techniques

Page 136

The favoured strategy is to focus on resolving the root causes, commencing with the primary root cause first, as this will eliminate most of the effects. However, depending on the nature of the required solution, we may need to take shorter term measures aimed a coping with the root effects ie. dealing with the symptoms. The relations diagram highlights both. When it comes to formulating strategies for dealing with root causes (or effects) the systematic diagram is a useful tool for generating the actions required to resolve them. This is discussed in the next section.

Problem Solving and Process Improvement Tools and Techniques

Page 137

The Systematic Diagram (Tree Diagram) What is a Systematic Diagram?

The systematic diagram is used to illustrate the means or activities needed to achieve specific goals and objectives. It provides an overview of the cascade of objectives and subobjectives through to the specific tasks and activities required to achieve them. It is a graphical representation of the different levels of objectives and tasks required to accomplish a broader goal which illustrates the logical relationships between each level and the tasks which must be achieved.

When do you use a Systematic Diagram?

The systematic diagram is most useful when broad goals or objectives need to be broken down into executable tasks, or complex tasks need to be broken down into their component parts. Difficulties with implementing goals and objectives often arise in situations where it may not be clear what exactly needs to be done in order to achieve the objectives. The systematic diagram allows the team to resolve such issues by providing a structured approach for working through objectives and translating them into actionable tasks.

How do you use a Systematic Diagram?

State and record the problem, goal or objective. The goal or objective may come from a strategic plan which is to be implemented or as a result of a factor identified using the relations and affinity diagram techniques. Write down the problem, goal or objective on the far left of a large sheet of paper and draw a box around it, as shown in Figure 7.10. This defines the primary objective.

I m p r o v e d e l i v e r y p e r f o r m a n c e
Figure 7.12. The Goal Statement - "Improve Delivery Performance"

Problem Solving and Process Improvement Tools and Techniques

Page 138

Generate the next level of the diagram. To generate the next level of the diagram, ask the question How will this objective be achieved. The answers to the question how will either define tasks to be actioned or other objectives, the achievement of which will fulfil the primary objective. In our example, using the results of the relations diagram as an input to the process, the answer to the question How do we improve delivery performance generates the secondary objectives shown in Figure 7.11.
S e c o n d a r y O b j e c t i v e s
E n s u r e a l l n e c e s s a r y i n f o r m a t i o n c a p t u r e d o n o r d e r

H o w ?

P r i m a r y O b j e c t i v e
I m p r o v e d e l i v e r y p e r f o r m a n c e

M a k e c r e d i t c h e c k p r o c e s s f a s t e r

E l i m i n a t e p i c k i n g d e l a y s

I m p r o v e t e c h n i c i a n s c h e d u l i n g

Figure 7.13. Ask the question "How?" to generate the next level Complete the diagram Continue building the systemic diagram until all branches have been completed. To do this, continue asking the question How? for each objective. When all the answers for a particular objective represent actionable tasks, then the branch can be considered completed. Continue until all the branches lead to actionable tasks. A completed systematic diagram is shown in Figure 7.12. Note: not all the branches will extend neatly together or generate tasks at the same level. Some objectives will generate both tasks and sub-objectives. These
Problem Solving and Process Improvement Tools and Techniques

Page 139

sub-objectives must then be followed through until all branches end in actionable tasks.
Sub-objectives and tasks Secondary Objectives
Ensure all necessary information captured on order Redesign orders proforma to ensure all information captured Modify computer input screen to check for all required information Change procedure to ensure incomplete orders cannot to approved Change procedure to bypass credit check for prequalified customers Introduce online credit check facility with bank

Primary Objective
Improve delivery performance

Make credit check process faster

Tasks
Install online link to bank for online credit check Install terminal in warehouse to allow daily check on due dates

Eliminate picking delays

Provide warehouse with advance notice of orders in the system Change procedure to send out of stock notification to Sales rep. Provide copy of daily work list to Sales Dept.

Objectives Actionable tasks

Improve technician scheduling

Give technicians advance notice of work that is coming up

Provide terminal for technicians to review new orders received

Figure 7.14. The completed systematic diagram Figure 7.12 shows a systematic diagram which has been completed within four levels. Every primary objective or problem will have its own unique hierarchy of objectives which may extend for considerably more than four levels. Do not be misled by the simplicity of the above example. It is important to follow through each branch however many levels it takes until all the actionable tasks which represent the final level are identified for each branch. These actionable tasks are also referred to as the fundamental means of a systematic diagram. The relationship between the various levels in the systematic diagram is defined by the questions How? and Why? illustrating the Ends and Means relationship between each two levels. This is illustrated in Figure 7.13.

Problem Solving and Process Improvement Tools and Techniques

Page 140

e a n s E n d s M
R e d e s i g n o r d e r s p r o f o r m a t o e n s u r e a l l i n f o r m a t i o n c a p t u r e d E n s u r e a l ln e c e s s a r y i n f o r m a t i o n c a p t u r e d o n o r d e r M o d i f y c o m p u t e r i n p u t s c r e e n t o c h e c k f o r a l l r e q u i r e d i n f o r m a t i o n C h a n g e p r o c e d u r e t o e n s u r e i n c o m p l e t e o r d e r s c a n n o t t o a p p r o v e d

H o w ?

H o w ?

H o w ?
M a k e c r e d i t c h e c k p r o c e s s f a s t e r C h a n g e p r o c e d u r e t o b y p a s s c r e d i t c h e c k f o r p r e q u a l i f i e d c u s t o m e r s I n t r o d u c e o n l i n e c r e d i t c h e c k f a c i l i t y w i t h b a n k I n s t a l l o n l i n e l i n k t o b a n k f o r o n l i n e c r e d i t c h e c k I n s t a l l t e r m i n a l i n w a r e h o u s e t o a l l o w d a i l y c h e c k o n d u e d a t e s

I m p r o v e d e l i v e r y p e r f o r m a n c e

W h y ?
E l i m i n a t e p i c k i n g d e l a y s P r o v i d e w a r e h o u s e w i t h a d v a n c e n o t i c e o f o r d e r s i n t h e s y s t e m C h a n g e p r o c e d u r e t o s e n d o u t o f s t o c k n o t i f i c a t i o n t o S a l e s r e p .

W h y ?

W h y ?
P r o v i d e c o p y o f d a i l y w o r k l i s t t o S a l e s D e p t . I m p r o v e t e c h n i c i a n s c h e d u l i n g G i v e t e c h n i c i a n s a d v a n c e n o t i c e o f w o r k t h a t i s c o m i n g u p P r o v i d e t e r m i n a l f o r t e c h n i c i a n s t o r e v i e w n e w o r d e r s r e c e i v e d

Figure 7.15. The Ends-Means relationship between the levels Assess the systematic diagram To understand the Ends-Means relationships illustrated by the systematic diagram: Start with any objective box in the diagram and ask the question How? to move down the hierarchy, ie. move to the right or ask the question Why? to move up the hierarchy, ie. move to the left. Start with any task box in the diagram and ask the question Why? to move up the hierarchy, ie. move to the left. Asking How? identifies the Means asking Why identifies the Ends. A completed systematic diagram will identify all the tasks which must be actioned in order to achieve all of the objectives identified in the diagram and in turn achieve the primary objective. Assess the results of the diagram to ensure that completion of the tasks identified do indeed achieve the higher level objective from which they were generated. If they do not, reassess the objective to see if you have missed some tasks or sub-objectives. Responsibility for completion of the action items is then assigned to team members.
Problem Solving and Process Improvement Tools and Techniques

Page 141

The Matrix Diagram What is a Matrix Diagram?

The matrix diagram is used to illustrate the relationship or interaction between two or more sets of factors. There are many variations on the basic matrix diagram including the: decision matrix; allocation matrix; prioritisation matrix.

This section shall consider the basic matrix diagram.

When do you use a Matrix Diagram?

The basic matrix diagram is most useful when you wish to assess the interaction or impact changes to a series of factors may have on a groups of other factors or areas. This is best illustrated by way of an example. which follows.

How do you use a Matrix Diagram?

Example: An organisation wishes to implement changes to their order fulfilment process. There are a number of planned changes including changes to written and computerised procedures, changes to responsibilities and job descriptions across several departments. In order to facilitate effective training and implementation in each group, the implementation team has decided to assess the degree to which each major group involved in the order fulfilment process is affected by each change, to ensure they are made aware of what they specifically will need to know and do differently. List the first group of factors along one of the matrix axes. In our case, the team have decided to list the changes to be made along the vertical axis of the matrix. This is illustrated in Figure 7.14..

Problem Solving and Process Improvement Tools and Techniques

Page 142

Changes New order input screen Authority tables change Online credit check facility New credit update screen Warehouse pick list update screen New orders list Technician backlog report Technician work allocations screen Figure 7.14 Matrix Diagram showing changes to be implemented List the second group of factors along the other matrix axis. In our example, the departments involved in the order fulfilment process have been listed along the horizontal axis. ie. along the top of the axis. This shown in Figure 7.15. Groups affected >> Changes New order input screen Authority tables change Online credit check facility New credit update screen Warehouse pick list update screen New orders list Technician backlog report Technician work allocations screen Figure 7.15 Matrix Diagram showing departments affected. Rate the level of interaction between the two groups of factors. Usually some form of rating system is used to indicate the level or type of interaction between the two groups of factors where some interaction exists. Sales Credit W/Hs Desp. Techs

Problem Solving and Process Improvement Tools and Techniques

Page 143

In our case, the team decided to use a system based on the letters: H = High, indicating the change has a high level of impact on this department; M = Moderate, indicating the change will have a moderate level of impact on this department; L = Low, indicating the change will have a low level of impact on this department;

Blank, indicating the change will have no impact on this department, The completed matrix diagram is shown in Figure 7.16. Use the results of the Matrix Diagram to guide future actions. In our example, the team would use the knowledge of which departments are most affected by which changes to develop specific implementation strategies, training and awareness sessions, focussed on the specific changes the different departments will need to master. For example, the implementation and training for the Technicians Department would focus on ensuring they were aware of the changes to authority tables and had access to the warehouse pick list. It would then focus on ensuring they had specific knowledge and skill in the use of procedures relating to, new order lists, backlog reporting and the work allocation screen. Alternately, the Despatch Department would only need to be aware of the authority table changes and have a moderate working knowledge of the warehouse picklist screen. Figure 7.16 The completed Matrix Diagram Groups affected >> Changes New order input screen Authority tables change Online credit check facility New credit update screen Warehouse pick list update screen New orders list Technician backlog report Technician work allocations screen Sales H M M L H H H Credit H H H W/Hs Desp. L L Techs L

H H

L M H M

Problem Solving and Process Improvement Tools and Techniques

Page 144

The Decision Matrix What is a Decision Matrix?

The Decision Matrix is a worksheet made up of columns and rows which is used to rank the options or alternatives which must be decided between, against a set of agreed criteria which the decision must meet, so as to facilitate decision making. It is best used by a team to arrive at a group decision in a manner that overcomes potential influences from personal biases of group members.

When do you use a Decision Matrix?

The decision matrix is most useful when a choice must be made between more than one clear alternative, and there are set of definable criteria which the decision must satisfy.

How do you use a Decision Matrix?

Determine the alternatives The first step before using a decision matrix is to determine the alternatives to be decided between. Consider an example of an organisation that wishes to progress a performance improvement initiative in their operations division. The alternatives have been defined as: Use an external consultant to progress the initiative; Use and internal team, trained and facilitated by an external consultant; Use an internal team only.

Determine the criteria to be satisfied by the ideal outcome. The criteria the successful course of action must meet, against which the alternatives will be evaluated against have been defined as: Implementation must be able to be fast tracked. A quick result is required; There must be a high degree of ownership of any changes by the employees; The results need to be sustainable in the long term; Implementation of recommendations should be easy;

Problem Solving and Process Improvement Tools and Techniques

Page 145

The methodology/approach should reflect leading edge management thinking; The project should be able to be done for the lowest cost possible; The conducting the project should not disrupt existing operations.

Lay out the decision matrix. Lay out the decision matrix with the alternatives along the horizontal axis and the criteria along the vertical axis. This is illustrated in Figure 7.17.
A l t e r n a t i v e s
E v a l u a t i o n C r i t e r i a U s e e x t e r n a l n t e r n a l t e a m C o n s u l t a n t s l e dI c o n s u l t a n t s i n l y n t e r n a l t e a m o

W e i g h t i n g S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e S c o r e S c o r e

F a s t t r a c k i m p l e m e n t a t i o n O w n e r s h i p o f a n y r e c o m m e n d e d c h a n g e s S u s t a i n a b l e r e s u l t s E a s e o f i m p l e m e n t a t i o n L e a d i n g e d g e a p p r o a c h L o w c o s t M i n i m i s e d i s r u p t i o n s t o e x i s t i n g o p e r a t i o n s T o t a l

Figure 7.17. Lay out the Decision Matrix Apply a weighting to each of the criteria. Assign a score to each of the criteria, which reflects the relative importance of that element. This is usually done by assigning a score out of 10 to each criteria. A Decision Matrix with the weightings applied to the criteria is shown in Figure 7.18.

Problem Solving and Process Improvement Tools and Techniques

Page 146

A l t e r n a t i v e s
E v a l u a t i o n C r i t e r i a U s e e x t e r n a l n t e r n a l t e a m C o n s u l t a n t s l e dI c o n s u l t a n t s i n l y n t e r n a l t e a m o

W e i g h t i n g S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e S c o r e S c o r e

F a s t t r a c k i m p l e m e n t a t i o n8 O w n e r s h i p o f a n y 4 r e c o m m e n d e d c h a n g e s S u s t a i n a b l e r e s u l t s 9 E a s e o f i m p l e m e n t a t i o n 4 L e a d i n g e d g e a p p r o a c h 2 L o w c o s t 3 M i n i m i s e d i s r u p t i o n s t o 7 e x i s t i n g o p e r a t i o n s T o t a l

Figure 7.18. Decision Matrix with each criteria weighted Score each alternative along each criteria. That is, assign a score out of ten for how each alternative performs along each element of the criteria. A scored decision matrix is shown in Figure 7.19.
A l t e r n a t i v e s
E v a l u a t i o n C r i t e r i a U s e e x t e r n a l n t e r n a l t e a m C o n s u l t a n t s l e dI c o n s u l t a n t s i n l y n t e r n a l t e a m o

W e i g h t i n g S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e S c o r e S c o r e

F a s t t r a c k i m p l e m e n t a t i o n8 O w n e r s h i p o f a n y 4 r e c o m m e n d e d c h a n g e s S u s t a i n a b l e r e s u l t s 9 E a s e o f i m p l e m e n t a t i o n 4 L e a d i n g e d g e a p p r o a c h 2 L o w c o s t 3 M i n i m i s e d i s r u p t i o n s t o 7 e x i s t i n g o p e r a t i o n s T o t a l

7 2 5 8 8 4 8

6 5 6 7 8 6 7

3 8 8 5 4 7 6

Figure 7.19. A scored Decision Matrix Determine a weighted score for each alternative Multiply the score by the weighting for each criteria for each alternative to obtain a weighted score. Add up the weighted scores to obtain a total for each alternative. This completes the decision matrix, shown in Figure 7.20.

Problem Solving and Process Improvement Tools and Techniques

Page 147

A l t e r n a t i v e s
E v a l u a t i o n C r i t e r i a U s e e x t e r n a l n t e r n a l t e a m C o n s u l t a n t s l e dI c o n s u l t a n t s i n l y n t e r n a l t e a m o

W e i g h t i n g S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e W e i g h t e d S c o r e S c o r e S c o r e

F a s t t r a c k i m p l e m e n t a t i o n8 O w n e r s h i p o f a n y 4 r e c o m m e n d e d c h a n g e s S u s t a i n a b l e r e s u l t s 9 E a s e o f i m p l e m e n t a t i o n 4 L e a d i n g e d g e a p p r o a c h 2 L o w c o s t 3 M i n i m i s e d i s r u p t i o n s t o 7 e x i s t i n g o p e r a t i o n s T o t a l

7 5 6 6 4 8 3 2 4 2 8 5 2 0 8 3 2 5 8 8 4 8 4 5 3 2 1 6 1 2 5 6 2 2 5 6 7 8 6 7 5 4 2 8 1 6 1 8 4 9 2 3 3 8 5 4 7 6 7 2 2 0 8 2 1 4 2 2 1 9

Figure 7.20. A completed Decision Matrix

Assess the results of the decision matrix and make the decision. Assess the result of the decision matrix. Ideally, the alternative which has resulted in the highest weighted total score will be the alternative which best meets the decision criteria.

Problem Solving and Process Improvement Tools and Techniques

Page 148

The Prioritisation Matrix What is a Prioritisation Matrix?

The Prioritisation Matrix is an application of the basic matrix diagram which is used to assign priorities to a group of activities needing to be implemented. Alternatively, this type of matrix may be used for evaluating and comparing a list of possible solutions against a set of guidelines, in the same way as a decision matrix is used. As with the other types of matrix based techniques, the prioritisation matrix is best used in the team environment, to arrive at a group decision in a manner that overcomes potential influences from personal biases of group members.

When do you use a Prioritisation Matrix?

A prioritisation matrix is most useful when a there are several countermeasures that may be implemented to solve a problem or several actions which may potentially be taken to implement a strategy, and where trade-offs need to be assessed between costs, benefits, ease of implementation, timing or similar characteristics. A typical example would be a situation where there may be a need to implement a quick fix to address the symptoms of a problem while progressing longer term actions aimed at resolving the root causes. A prioritisation matrix could be used to establish implementation priorities between the range of available actions.

How do you use a Prioritisation Matrix?

Determine the guidelines the countermeasures or actions should meet Determine the guidelines or criteria the countermeasures or actions to be implemented must meet. These will be the criteria against which the countermeasures or actions shall be evaluated. Consider the example of an organisation wishing to implement a range of measures aimed at improving the responsiveness of their order fulfilment process. To develop a detailed implementation strategy, the team need to prioritise the actions to establish the best order to implement so as to maximise the benefits and the speed with which they are achieved.

Problem Solving and Process Improvement Tools and Techniques

Page 149

The criteria or guidelines for prioritisation have been decided as: Cost; Time required to implement; Ease of implementation; Authorisation, the degree to which higher level authorisation will be required.

Lay out the prioritisation matrix. List the guidelines or criteria along the horizontal axis and the range of actions or decision options along the vertical axis. This is illustrated in Figure 7.21.
A c t i o n s o r O p t i o n s G u i d e l i n e s

A c t i o n s C o s t T i m e E a s eA u t h o r i s a t i o n T o t a l R e d e s i g n o r d e r s p r o f o r m a t o e n s u r e a l l i n f o r m a t i o n c a p t u r e d M o d i f y c o m p u t e r i n p u t s c r e e n t o c h e c k f o r a l l r e q u i r e d i n f o r m a t i o n C h a n g e p r o c e d u r e t o e n s u r e i n c o m p l e t e o r d e r s c a n n o t t o a p p r o v e d C h a n g e p r o c e d u r e t o b y p a s s c r e d i t c h e c k f o r p r e q u a l i f i e d c u s t o m e r s I n s t a l l o n l i n e l i n k t o b a n k f o r o n l i n e c r e d i t c h e c k C h a n g e p r o c e d u r e t o s e n d o u t o f s t o c k n o t i f i c a t i o n t o S a l e s r e p . P r o v i d e c o p y o f d a i l y w o r k l i s t t o S a l e s D e p t . I n s t a l l t e r m i n a l i n w a r e h o u s e t o a l l o w d a i l y c h e c k o n d u e d a t e s P r o v i d e t e r m i n a l f o r t e c h n i c i a n s t o r e v i e w n e w o r d e r s r e c e i v e d

Figure 7.21. Lay out for prioritisation matrix Score each action or option according to each criteria or guideline. Score each action or option according to each criteria or guideline. A typical systems is to assign a score out of 10, where the lower the score the higher the degree of fit between the action/option and the criteria. eg. activity one - redesign orders proforma to ensure all information captured represents a negligible cost to implement. With reference to Figure 7.22 which shows a completed scored matrix, it only scores a 8 for this criteria. Installing a direct online link to the bank, the fifth action on this list is the most expensive to implement. It scores and 8 on the dimension on cost.
Problem Solving and Process Improvement Tools and Techniques

Page 150

Add up the score in each criteria to give a total for each action or option. The scored and completed matrix is shown in Figure 7.22.
A c t i o n s o r O p t i o n s G u i d e l i n e s

A c t i o n s C o s t T i m e E a s eA u t h o r i s a t i o n T o t a l R e d e s i g n o r d e r s p r o f o r m a t o e n s u r e a l l 1 3 2 2 8 i n f o r m a t i o n c a p t u r e d M o d i f y c o m p u t e r i n p u t s c r e e n t o c h e c k f o r 3 5 2 5 1 5 a l l r e q u i r e d i n f o r m a t i o n C h a n g e p r o c e d u r e t o e n s u r e i n c o m p l e t e 2 2 3 7 1 4 o r d e r s c a n n o t t o a p p r o v e d C h a n g e p r o c e d u r e t o b y p a s s c r e d i t c h e c k 2 3 5 9 1 9 f o r p r e q u a l i f i e d c u s t o m e r s I n s t a l l o n l i n e l i n k t o b a n k f o r o n l i n e c r e d i t 8 7 3 9 2 7 c h e c k C h a n g e p r o c e d u r e t o s e n d o u t o f s t o c k 2 2 1 2 7 n o t i f i c a t i o n t o S a l e s r e p . P r o v i d e c o p y o f d a i l y w o r k l i s t t o S a l e s 2 1 1 2 6 D e p t . I n s t a l l t e r m i n a l i n w a r e h o u s e t o a l l o w d a i l y 5 5 1 7 1 8 c h e c k o n d u e d a t e s P r o v i d e t e r m i n a l f o r t e c h n i c i a n s t o r e v i e w 5 5 1 7 1 8 n e w o r d e r s r e c e i v e d

Figure 7.22. Each action scored for each criteria/guideline Use the resulting total scores to determine priorities. The general rule is to prioritise the actions in order of ascending score. That is, implement the lowest scoring option/action first. The completed and prioritised matrix is shown in Figure 7.23.
A c t i o n s o r O p t i o n s G u i d e l i n e s
I m p l e m e n t a t i o n p r i o r i t i e s

A c t i o n s C o s t T i m e E a s e A u t h o r i s a t i o n T o t a l R e d e s i g n o r d e r s p r o f o r m a t o e n s u r e a l l 1 3 2 2 8 i n f o r m a t i o n c a p t u r e d M o d i f y c o m p u t e r i n p u t s c r e e n t o c h e c k f o r 3 5 2 5 1 5 a l l r e q u i r e d i n f o r m a t i o n C h a n g e p r o c e d u r e t o e n s u r e i n c o m p l e t e 2 2 3 7 1 4 o r d e r s c a n n o t t o a p p r o v e d C h a n g e p r o c e d u r e t o b y p a s s c r e d i t c h e c k 2 3 5 9 1 9 f o r p r e q u a l i f i e d c u s t o m e r s I n s t a l l o n l i n e l i n k t o b a n k f o r o n l i n e c r e d i t 8 7 3 9 2 7 c h e c k C h a n g e p r o c e d u r e t o s e n d o u t o f s t o c k 2 2 1 2 7 n o t i f i c a t i o n t o S a l e s r e p . P r o v i d e c o p y o f d a i l y w o r k l i s t t o S a l e s 2 1 1 2 6 D e p t . I n s t a l l t e r m i n a l i n w a r e h o u s e t o a l l o w d a i l y 5 5 1 7 1 8 c h e c k o n d u e d a t e s P r o v i d e t e r m i n a l f o r t e c h n i c i a n s t o r e v i e w 5 5 1 7 1 8 n e w o r d e r s r e c e i v e d

3 5 4 8 9 2 1 6 7

Figure 7.23. The completed and prioritised matrix

Problem Solving and Process Improvement Tools and Techniques

Page 151

The Allocation Matrix What is a Allocation Matrix?

The Allocation Matrix is usually used as an extension to the systematic diagram to allocate responsibilities for actioning the various tasks that are identified as a result of applying the systematic diagram tool to an objective.

When do you use an Allocation Matrix?

An allocation matrix is most useful when there are many tasks to be completed in order to achieve some higher order goal/objective, and tasks must be allocated to individual team members for completion in a coordinated fashion.

How do you use an Allocation Matrix?

Layout the matrix Commence with the systematic diagram. At the far right of the systematic diagram, once all the required tasks or actions have been identified, draw a matrix consisting of three columns and a row for each task identified by the systematic diagram. This is illustrated in Figure 7.24.
Who
Redesign orders proforma to ensure all information captured

When

Status

Ensure all necessary information captured on order

Modify computer input screen to check for all required information

Change procedure to ensure incomplete orders cannot to approved

Make credit check process faster Introduce online credit check facility with bank

Change procedure to bypass credit check for prequalified customers Install online link to bank for online credit check

Improve delivery performance

Eliminate picking delays

Provide warehouse with advance notice of orders in the system

Install terminal in warehouse to allow daily check on due dates Change procedure to send out of stock notification to Sales rep. Provide copy of daily work list to Sales Dept. Provide terminal for technicians to review new orders received

Improve technician scheduling Give technicians advance notice of work that is coming up

Figure 7.24. Lay out the allocation matrix

Problem Solving and Process Improvement Tools and Techniques

Page 152

Label the three columns Who, When and Status as shown in Figure 7.24. Assign tasks to team members. Assign tasks to team members by writing in the team member name in the Who column and specify completion date requirements in the When column. This is illustrated in Figure 7.25.
Who John Sue Peter Peter Rob
Introduce online credit check facility with bank Install online link to bank for online credit check

When End June End Sept End June End July Jan next year End Oct End July End June End Sept

Status

Redesign orders proforma to ensure all information captured

Ensure all necessary information captured on order

Modify computer input screen to check for all required information

Change procedure to ensure incomplete orders cannot to approved

Make credit check process faster

Change procedure to bypass credit check for prequalified customers

Improve delivery performance

Eliminate picking delays

Provide warehouse with advance notice of orders in the system

Install terminal in warehouse to allow daily check on due dates Change procedure to send out of stock notification to Sales rep. Provide copy of daily work list to Sales Dept. Provide terminal for technicians to review new orders received

Sue John Deb Sue

Improve technician scheduling Give technicians advance notice of work that is coming up

Figure 7.25. The completed allocation matrix Use the allocation matrix to monitor progress. Use the allocation matrix to monitor progress towards the achievement of the overall goal/objective. Use the status column to check off completed tasks. This is illustrated in Figure 7.26.

Problem Solving and Process Improvement Tools and Techniques

Page 153

Who John Sue Peter Peter Rob

Introduce online credit check facility with bank Install online link to bank for online credit check

When End June End Sept End June End July Jan next year End Oct End July End June End Sept

Status

Redesign orders proforma to ensure all information captured

Ensure all necessary information captured on order

Modify computer input screen to check for all required information

Change procedure to ensure incomplete orders cannot to approved

Make credit check process faster

Change procedure to bypass credit check for prequalified customers

Improve delivery performance

Eliminate picking delays

Provide warehouse with advance notice of orders in the system

Install terminal in warehouse to allow daily check on due dates Change procedure to send out of stock notification to Sales rep. Provide copy of daily work list to Sales Dept. Provide terminal for technicians to review new orders received

Sue John Deb Sue

Improve technician scheduling Give technicians advance notice of work that is coming up

Figure 7.26. Using the allocation matrix to track progress

Problem Solving and Process Improvement Tools and Techniques

Page 154

The Process Decision Program Chart What is a Process Decision Program Chart?
The Process Decision Program Chart (PDPC) is a tool which helps a team anticipate and plan for dealing with problems which may arise during implementation. It identifies likely weaknesses and potential for failure. The team can then use the PDPC to formulate and assess contingency plans and countermeasures to minimise the risk of failure. The Failure Modes and Effects Analysis Technique (FMEA) often used in conjunction with new product development is based on the use of the PDPC tool.

When do you use a Process Decision Program Chart?

The PDPC is best used to help a team with the development of contingency plans. It is also useful as a tool for evaluating alternative courses of action.

How do you use a Process Decision Program Chart?

Identify the goal or objective to be achieved. Consider an organisation whose objective is to ensure all customer orders are delivered on time. The objective can be defined as on-time order delivery. Brainstorm all the possible unexpected problems Identify all the possible unexpected problems which could occur in each part of the process associated with filling customer orders. Ask the question What if ... ?. Figure 7.27 illustrates a Process Decision Program Chart showing the things which might go wrong and act against the achievement of the objective ontime order delivery.

Problem Solving and Process Improvement Tools and Techniques

Page 155

P r o m i s e d d e l i v e r y t i m e t o o s h o r t S a l e s a n d o r d e r r e c e i p t N o t e n o u g h i n f o r m a t i o n p r o v i d e d t o f i l l o r d e r

C u s t o m e r f a i l s c r e d i t c h e c k C r e d i t c h e c k N o r e s p o n s e f r o m b a n k

O n t i m e o r d e r d e l i v e r y

I t e m n o t i n s t o c k W a r e h o u s e p i c k i n g N o t i f i c a t i o n o f r e q u i r e m e n t s n o t r e c e i v e d i n t i m e

C u s t o m e r d e l i v e r i e s

S u i t a b l e t r a n s p o r t n o t a v a i l a b l e w h e n r e q u i r e d

Figure 7.27. PDPC with "what if's" identified Formulate contingencies and countermeasures. Develop contingency plans and countermeasures which can be used to overcome the effect of the what ifs so as to still achieve the objective of on-time order delivery. This is illustrated in Figure 7.28.
P r o m i s e d d e l i v e r y t i m e t o o s h o r t S a l e s a n d o r d e r r e c e i p t N o t e n o u g h i n f o r m a t i o n p r o v i d e d t o f i l l o r d e r
C h e c k w i t h l o g i s t i c s b e f o r e p r o m i s i n g d e l i v e r y t i m e s R e d e s i g n p r o f o r m a t o e n s u r e a l l i n f o c a p t u r e d

C u s t o m e r f a i l s c r e d i t c h e c k C r e d i t c h e c k N o r e s p o n s e f r o m b a n k

S u p p l y a n y w a y

O n t i m e o r d e r d e l i v e r y

R e c h e c k a l l a c c o u n t s i f n o r e s p o n s e f o r 2 h o u r s R a i s e p u r c h a s e o r d e r P r o v i d e o n l i n e a c c e s s t o n e w o r d e r s

I t e m n o t i n s t o c k W a r e h o u s e p i c k i n g N o t i f i c a t i o n o f r e q u i r e m e n t s n o t r e c e i v e d i n t i m e

C u s t o m e r d e l i v e r i e s

S u i t a b l e t r a n s p o r t n o t a v a i l a b l e w h e n r e q u i r e d

U s e t h i r d p a r t y c o u r i e r

Figure 7.28. Countermeasures/contigencies identified for "what ifs"

Problem Solving and Process Improvement Tools and Techniques

Page 156

The countermeasures or contigencies are indicated in the ovals to distinguish them from the what ifs represented in boxes. Assess the PDPC Assess the viability of implementing the identified countermeasures and contingencies. This is illustrated in Figure 7.29.
P r o m i s e d d e l i v e r y t i m e t o o s h o r t S a l e s a n d o r d e r r e c e i p t N o t e n o u g h i n f o r m a t i o n p r o v i d e d t o f i l l o r d e r
C h e c k w i t h l o g i s t i c s b e f o r e p r o m i s i n g d e l i v e r y t i m e s R e d e s i g n p r o f o r m a t o e n s u r e a l l i n f o c a p t u r e d

C u s t o m e r f a i l s c r e d i t c h e c k C r e d i t c h e c k N o r e s p o n s e f r o m b a n k

S u p p l y a n y w a y

O n t i m e o r d e r d e l i v e r y

R e c h e c k a l l a c c o u n t s i f n o r e s p o n s e f o r 2 h o u r s R a i s e p u r c h a s e o r d e r P r o v i d e o n l i n e a c c e s s t o n e w o r d e r s

I t e m n o t i n s t o c k W a r e h o u s e p i c k i n g N o t i f i c a t i o n o f r e q u i r e m e n t s n o t r e c e i v e d i n t i m e

C u s t o m e r d e l i v e r i e s

S u i t a b l e t r a n s p o r t n o t a v a i l a b l e w h e n r e q u i r e d

U s e t h i r d p a r t y c o u r i e r

R e c o m m e n d e d a c t i o n D i f f i c u l t o r I m p o s s i b l e

Figure 7.29. The completed PDPC

Problem Solving and Process Improvement Tools and Techniques

Page 157

The Network Diagram (Critical Path) What is a Network Diagram?

A Network Diagram is a technique used for the planning and tracking of projects. It is a technique which identifies the critical path of a project, which represents the shortest possible time in which the project may be completed. It also forms the basis for the development of bar charts or gantt charts for monitoring and controlling projects.

When do you use a Network Diagram?

The network diagram is best used whenever there is a requirement to implement a project that uses significant resources and requires the time phased coordination of several types of skills to complete. It is also particularly useful in situations where specific completion dates must be met and to identify the available options and flexibility for reducing time to completion and optimising the use of resources.

How do you use a Network Diagram?

The use of the network diagram technique, known as critical path planning is outside the scope of this handbook.

Problem Solving and Process Improvement Tools and Techniques

Page 158

Appendix: Understanding Data and Variation

Problem Solving and Process Improvement Tools and Techniques

Page 159

1. Introduction to Variation
One of the laws of nature is that no two things are exactly identical. No two manufactured items are exactly the same, no two services are delivered in exactly the same manner. Variation, however small, is a fact of life and of business. It is therefore important to understand variation so that we may manage it proactively to our benefit, and not be continually reacting in order to compensate for the impact it may be having on our processes and on our business.

2. Observing Variation
If we collect data on any process and plot the data values on a histogram, we can observe the variation in the process. The distribution of points plotted will give rise to a pattern that characterises the observed variation. Some typical patterns that are often formed by different processes are illustrated in Figure 2.1.

3. Characterising Data
For a given group of data, we can determine several characteristics that allow us to describe and characterise the data and its variation. The key measures that allow us to do this are measures of: Shape Spread Location

These are illustrated in Figure 2.2.

Problem Solving and Process Improvement Tools and Techniques

Page 160

Normal Distribution

Skewed or "cliff like" distribution

"Comb like" distribution

Bi-modal distribution ie. two peaks

Normal with outlying data

Figure 2.1 Examples of Variation Patterns

Problem Solving and Process Improvement Tools and Techniques

Page 161

L o c a t i o n

L o c a t i o n

S p r e a d

S p r e a d

S h a p e N o r m a l

S h a p e S k e w e d

Figure 2.2 Characterising Data

Problem Solving and Process Improvement Tools and Techniques

Page 162

3.1 Measures of Location

There are three measures of location used in statistics. Mean _ X or the mean is the ordinary arithmetic average of a group of numbers. It is referred to as X-bar when spoken. _ The mean X = Xn n Where Xn Xn n = = = represents the values in the sample of data the sum of all the values in the sample the number of values in the sample (sample size)

Example: Suppose an organisation collects some data on the performance of its order fulfilment process. The number of days it had taken to fill eleven consecutive orders was recorded as: 9, 6, 4, 2, 4, 2, 4, 7, 2, 4, 3. The mean or average time to fill an order, based on this sample is calculated as: _ X = Xn = 9 + 6 + 4 + 3 + 4 + 2 +5 + 7 + 2 + 6 + 3 n 11 = 51 11 = 4.6 days

The mean time to fill an order is 4.6 days.

Problem Solving and Process Improvement Tools and Techniques

Page 163

The Median The median is the middle value of a group of data points. It is the value such that half the data points are less than the median and half are greater. If data is skewed or there are a few extreme values, these will distort the calculation of the mean. The result being, the value obtained for the mean will not be typical of most of the values in the data. ie. the value of the mean will not be where most of the data points are located. For example: the median is often used when reporting housing prices, because the occasional sale of a very expensive house distorts the average calculation. Prospective home buyers, are more likely to be interested in what the typical house in a particular area is likely to sell for. In such cases, the mean will not be a good indicator of this. Consider the following data, showing the values of houses sold in a particular suburb during a month. Five houses were sold, for the following amounts: $ 130,000, $ 132,000, $ 420,000, $ 137,000, $ 135,000 The mean housing price from this data is: $130,000 + $132,000 + $420,000 + $137,000+$135,000 5 = $ 190,800

We can see from observation, that with the exception of the $420,000 house, the others are all in the range between $130,000 and $140,000. The prospective home buyer wanting to buy a typical home in this suburb would therefore need to budget for somewhere in this range. The mean value of $190,800 has been distorted by the single $420,000 home which was not typical of the others. To prospective home buyers, this would be a misleading representation of housing values in this suburb. To calculate the median of the above data, we re-list the values from the lowest to the highest value ie. ascending order. eg.: $130,000, $ 132,000, $135,000, $137,000, $420,000 The middle value is the third, or $ 135,000. Therefore, we can say that the median house price in this suburb during the month the data was collected was
Problem Solving and Process Improvement Tools and Techniques

Page 164

$135,000. We can see that this is indeed typical of what most of the houses sold for, and has overcome the distorting effect of the single $420,000 house. Therefore, in situations where the data is skewed or there are extreme values present, the median is a much better indicator of location than the mean. Let use reconsider the performance of the order fulfilment process, for which we have previously calculated the mean. 9, 6, 4, 2, 4, 2, 4, 7, 2, 4, 3. To find the median of this data, we reorder it from the lowest to the highest value. ie. 2, 2, 2, 3, 4, 4, 4, 4, 6, 7, 9 There are 11 values in our group of data points, the middle value will be the 6th. ie. their are 5 values below it, 5 values above it. The 6th value is 4 days. Therefore we can say the median = 4 days. In cases where there are an even number of data points, the median is calculated as the sum of the middle two divided by 2. Suppose we had ten data points as follows: 2, 2, 2, 3, 4, 5, 5, 6, 6, 7 There is no middle value. The two middle values, the 5th and 6th are 4 and 5 respectively. The median is then calculated as follows: Median = 4+5 = 2 9 2 = 4.5

Problem Solving and Process Improvement Tools and Techniques

Page 165

The Mode The mode is defined as the most frequently occurring value. Again, in situations where skewed data, or the presence of extreme values give a false impression of the location of the majority of the data, the mode provides a more useful measure of location by indicating which is the most frequently occurring data point. With reference to our original set of data on the performance of the order fulfilment process: 9, 6, 4, 2, 4, 2, 4, 7, 2, 4, 3. The most frequently occurring value, which occurs four times is the 4. Therefore the mode = 4.

Problem Solving and Process Improvement Tools and Techniques

Page 166

3.2 Measures of Spread

Range The range indicates the spread of extreme values. It is calculated by subtracting the smallest value in the sample from the largest and is represented by the symbol R. With reference to the data referred to in our previous examples. 9, 6, 4, 2, 4, 2, 4, 7, 2, 4, 3. The smallest value = 2, the largest value = 9. The Range can be calculated as: R = 9-2 = 7 days.

The Standard Deviation The range is an easy way to get an indication of the amount of variation, but it does not take into account the overall spread of values. Figures 3.1a to 3.1c show three different distributions of data.
R a n g e

M e a n

Figure 3.1.a Data points "tightly" clustered around the mean

Problem Solving and Process Improvement Tools and Techniques

Page 167

R a n g e

M e a n

Figure 3.1.b. Data points evenly spread across the range

R a n g e

M e a n
Figure 3.1.c. Data points clustered around the range extremes Depending on the values of specific data points, all three distributions could conceivably have the same values for the mean, median and range. Yet, they are very different from each other and would indicate significantly different process performance situations. The standard deviation, often referred to as SD, S, or the Greek symbol sigma is a measure which overcomes this problem, and allows us to develop a more complete picture of what our data it telling us. The standard deviation is a measure of how tightly clustered around the mean, our data is. It is a measure which reflects the variability and dispersion of data. It is effectively a measure of the average distance from the mean of all of the data points. Therefore, data which is tightly clustered around the mean will
Problem Solving and Process Improvement Tools and Techniques

Page 168

have a relatively low value for standard deviation compared to data which is dispersed away from the mean. With reference to Figure 3.1a to 3.1c. Figure 3.1.a would have the smallest standard deviation - most points clustered close to the mean, Figure 3.3.c. would give the highest value - most number of points relatively far from the mean. The standard deviation is calculated as follows. _ s = ( Xn - X )2 n-1 _ The term ( Xn - X )2 is referred to as the Variance (Var or S2 ). n-1 The standard variation is the square root of the variance. The term Xn refers to our data points; n is the number of data points we have; X is the mean of the data.

With reference to our sample data about a companys order fulfilment process, we can calculate the standard deviation as follows. 9, 6, 4, 2, 4, 2, 4, 7, 2, 4, 3 We know from our previous calculation, that the mean is 4.6 days. Variance = (9-4.6)2 + (6-4.6)2 + (4-4.6)2 + ...... + (3-4.6)2 (11 - 1)

19.36 + 1.96 + 0.36 + ......... + 2.56 10 51.36 10 = 5.14 = = = 2.27 Var 5.14

Variance

Standard Deviation Standard Deviation Standard Deviation

Problem Solving and Process Improvement Tools and Techniques

Page 169

4. Understanding Variation 4.1 Common and Special Causes

Irrespective of what individual data points are produced by a process, if the shape, location and spread of the data remains consistent, then we are able to predict how the process will perform. This allows us to make decisions that take into account the variation which we now is there, in this way minimising the negative impact it may potentially have on the operation of our processes. If the shape, location and spread is forever changing, then we are unable to predict how a process with perform, giving rise to uncertainty in the decision making process and in the way we manage our business. An American statistician, Dr. W.A. Shewhart, studied variation and patterns of variation during the 1920's. The results of his studies yielded a major breakthrough in the understanding of variation and how it should be managed. The results of his studies can be summarised by three simple statements: All processes display variation; Some display controlled variation; Some display uncontrolled variation.

Another American statistician Dr. W. Edwards Deming, renamed these two different types of variation by calling them variation due to special causes and variation due to common causes. Variation due to common causes was due to the many potential variables interacting in different combinations as part of the normal operation of a system or process. This type of variation was purely random and people working in a system have no control over it. Measuring and plotting data from a process which exhibited common causes only would produce distributions with consistent shape, location and spread. Such processes are said to be in a controlled state, stable or in statistical control. Variation due to special causes was due to causes which were not a natural part of the process or system. This type of variation could always be traced to an specific or "assignable" cause, such as an unusual event, a specific person, machine or localised condition.

Problem Solving and Process Improvement Tools and Techniques

Page 170

Data collected at different times from processes or systems in which special causes are present, will produce distributions which are not consistent as far as shape, location and spread are concerned. Such processes are said to be uncontrolled, unstable or not in statistical control. Shewhart's breakthrough pointed the way to quality and productivity improvement through the reduction of variation by removing special causes first, to make the process stable, and then working on the common causes to reduce the variation inherent in the process. Deming went further with his definition of quality as: Good quality does not necessarily mean high quality. It mean a predictable degree of uniformity and dependability at low cost, with a quality suited to the market. To Deming, quality meant reducing variation, so that customer requirements could be consistently met. In addition, Deming highlighted the fact that because common causes "belong" to a system or process, and are a fundamental property of a process, people working in the process had no control over them. It was pointless blaming employees for poor performance. If the substandard performance was within the limits of the common cause variation of the process they were working in, employees were effectively "prisoners of the process". Equally pointless was rewarding managers for above standard performance if this too was within the limits of natural variation of the process. Both these results are due to the process and not as a result of the actions of any individual on the process. To remove common causes requires the implementation of fundamental changes to the process or system which only management have the power to change. Special causes on the other hand, can be traced to an individual, machine, event or local condition. Employees and supervisors working in a process or system can usually act to identify and remove special causes. Therefore, for an organisation to work towards continuous improvement, it needs to train and empower employees to identify and remove special causes from their work processes until a state of statistical control was achieved. When processes have been stabilised and only common cause variation exists, management can either themselves, or through the use of empowered
Problem Solving and Process Improvement Tools and Techniques

Page 171

employee teams, work on changing or completely re-engineering the business processes to address and reduce common cause variation, and to reposition the location of the process, if necessary, to more accurately target the customer requirements. The issue for management, is understanding the variation inherent in a process, and being able to tell the difference between common cause variation and special cause variation. Shewharts main contribution to the understanding and management of variation was to provide a tool which management could use to both, monitor variation and discriminate between common cause and special cause variation. This tool was the control chart. The control chart is based on the Normal distribution, illustrated in Figure 4.1.

Average = Median

2s = 68.26 % of Data

4s = 95.44 % of Data 6s = 99.73 % of Data

Figure 4.1 The Normal Distribution The Normal distribution has some very useful properties. Namely: 68.26% of all data values will fall within the band of plus or minus one standard deviation around the mean;

Problem Solving and Process Improvement Tools and Techniques

Page 172

95.44 % of all data values will fall within the band of plus or minus two standard deviations around the mean; 99.73 % of all data values will fall within the band of plus or minus three standard deviations around the mean.

If we have a process or system, whose performance generates data that is normally distributed, we can use our data to calculate and draw an upper control limit three standard deviations above the mean and a lower control limit three standard deviations below the mean, as illustrated in Figure 4.2. The range spanned by these limits will enclose 99.73% of all the data points that are likely to be produced by the process or system.

U p p e r C o n t r o l L i m i t = X + 3 x S D

9 9 . 7 3 % o f a l l d a t a p o i n t s

M e a n = X

L o w e r C o n t r o l L i m i t = X 3 x S D

Figure 4.2 A Control Chart Layout These limits would effectively define the limits of common cause variation attributable to a system (ie. covering 99.73% of all possible outcomes). Data points which lying outside these limits, will usually be as a result of special causes. Such a chart, showing the upper and lower control limits, the mean and plotting the data points is known as a control chart. Fortunately, most naturally occurring data in business and in the natural world tends to follow the normal distribution, or follows distributions which for practical purposes may be approximated to normal. Where this is not the case, the Normal distribution can be approximated with appropriate sampling procedures. In addition, when data collection and analysis is based on samples of data and sample averages are taken as data points rather than the individual points, Shewhart proved that regardless what the original distribution was, the averages will approximate the Normal distribution.
Problem Solving and Process Improvement Tools and Techniques

Page 173

This tendency of averages of data to be more normally distributed than the original data is often referred to as the Central Limit Theorem. This means that the control chart and Shewharts principles in general have widespread application for understanding and managing variation.

4.2. Collecting Data

When you collect data you will usually be collecting a sample which you hope is representative of all the outcomes for a particular process. eg. a process may produce 1,000 manufactured items per day, or process 200 transactions per day. In most cases it will be impractical to examine each item or transaction, and so a sample is be taken. In general, the total number of outcomes eg. transactions, items etc is referred to as the population. The subgroup which is selected for examination is called a sample. The two basic approaches to sampling are: Random sampling, where sampling is carried out in such a way to ensure that every unit in a population has an equal chance of being included in a sample, regardless of the unit's appearance or position. Systematic sampling, where samples are taken at fixed intervals.

Because we are only looking at a small part of the total population with our sample, there will inevitably be some difference between the results obtained from our sample and the result we would have obtained had we looked at the whole population. The variation between the population and the sample results in sampling error or sampling variation. The topic of sampling is a complex subject in its own right and outside the scope of this booklet. Suffice to say, one needs to be aware of the impact of sampling error on calculations, so as not to jump to incorrect conclusions on the basis of small or limited samples. eg. small apparent improvements in a process may be due to sampling error effects and not due to any genuine improvement. In all critical situations, one should seek skilled advice on sampling and sample size selection.

Problem Solving and Process Improvement Tools and Techniques

Page 174

End of Tools and Techniques Guidebook

Problem Solving and Process Improvement Tools and Techniques

Page 175