WICKED PROBLEMS, RIGHTEOUS SOLUTIONS

BACK TO THE FUTURE ON LARGE COMPLEX

PROJECTS
Robert Lane
1
and Graham Woodman
2
ABSTRACT
Most large, long and complex projects are Wicked Problems. To deal with change and
uncertainty on such projects, the project Delivery Process (Definition, Design,
Manufacture and Assembly) needs to be considered with the Development Process
(Business case, statement of need, Functional Brief) as a total system. The Development
Process needs to deliver a minimum usable subset of Business, Customer and
Operational requirements to enable the Delivery process to start. In return the Delivery
Process needs to improve flexibility to allow the Business, Customer and Operator to
respond to changes due to technology improvements or market conditions and still
improve efficiency and be lean.
The project team for Terminal 5 at Heathrow Airport has identified tools and techniques
to help solve the wicked problems of Project Development and also to improve flexibility
of Delivery through the use of Last Responsible Moment (LRM) for information transfer
and decisions. These are established by working backwards from completion. The LRM
concept together with lean techniques and decoupled project delivery systems has given
a new view on traditional Project Management techniques and project processes. The
worlds most refreshing Interchange project of Terminal 5 at Heathrow costing some
1.9Billion is starting the delivery process utilising LRM concepts in an information
driven project system developed in-house using simple rules and a simple visual basic
programme which links process mapping, planning & programming and information
control
KEY WORDS
Last Responsible Moment, Information Driven Project System

1
Airside Civil Engineering Delivery Manager, BAA Terminal 5 Project Team, PO Box 620, Longford
House, 420 Bath Road, Longford, West Drayton, England UB7 0NX (44)208 745 1628 FAX (44)208
745 2379, [email protected]
2 Associate Director TPS Consult, The Landsdowne Building, Landsdowne Road, Croydon, CR0 2BX,
England, (44)208 256 4226, FAX (44)208 256 4572, [email protected]
INTRODUCTION
SOLVING THE RIGHT PROBLEMS
Successful problem solving requires finding the right solution to the right problem. We
fail more often because we solve the wrong problem than because we get the wrong
solution to the right problem Russell Ackoff 1974
We live in turbulent , fast moving times. The days have gone when certainty could be
measured for years ahead. The cycle of change winds ever tighter and businesses and
organisations need flexibility to respond to changes in market conditions and technology,
often played out on a global stage. Construction projects are generally only a means to an
end and form only one part of the response in the chain of identifying and meeting
business needs. Too often the construction industry has failed to respond to this basic
requirement of flexibility by demanding fixed requirements at the start of a project and
then complaining when the customer or owner changes his mind. Large, complex and
long projects suffer from requirements volatility. If there are comments and thoughts
such as Just tell me what you want and we will do it. or The customer keeps changing
their mind or Where is the brief? then there is probably an underlying wicked
problem (Rittel and Webber 1973).
For every complex problem there is a simple solution. And it is wrong. H.L.Menken
Wicked problems are often those where the dynamic and behaviourial complexities
are high; where different groups of key decision makers hold different assumptions,
values and beliefs, and where component problems cannot be solved in isolation from one
another. Conversely tame problems (Rittel and Webber 1973) have low dynamic and
behaviourial complexity and can be solved using conventional analytical methods
involving data collection and static analysis (i.e. analysis that does not require dealing
with delays, multiple feedback loops, and non-linear relationships). Tame problems can
be solved in isolation, can be broken down into parts which can be solved independently
by different groups of people. Solutions to different parts of a larger problems can then be
integrated into an overall solution. There is an old Japanese saying If all you have is a
hammer , then everything looks like a nail and trying to solve wicked problems using
tame problem solving techniques will cause the wrong problem to be solved.
AIRPORTS OF THE FUTURE
BAA plc is the worlds largest commercial operator of airports, operating seven UK
airports handling some 112 million passengers each year and all or part of eight other
airports in the rest of the world. Competition is played on the world stage. Airlines form
into global alliances with passengers offered routes not through adjacent UK airports but
via Paris, Schipol or Frankfurt. Providing capacity to meet air traffic growth is one of
BAAs primary duties. Terminal 5 at Heathrow was foreseen prior to BAA submitting a
Statement of Case in January 1995. The Public Inquiry into the proposal to build a fifth
terminal at Heathrow Airport started in May 1995 and finally finished after having sat for
525 days, in March1999, making it the longest public Inquiry in British planning history.
The Inspector is currently writing his report recommending whether the terminal should
go ahead. A Government decision is expected in 2001 with an anticipated start on site in
spring 2002 and an operational date of Spring 2007. Evidence offered at the Public
Inquiry to persuade the Inspector becomes effectively a constraint within the scheme and
is extremely difficult to change. Against this linear timeline background the airline and
airport businesses are subject to continuous change and pressure.
THE AIRPORTS WE DESIGN TODAY MUST BE FIT FOR THE WORLD TOMORROW.
What will it be like travelling through Heathrow Airport in the next century? Will there
still be queues? Will we travel at all, or just use virtual travel? BAA spends well over 1
million a day on building pieces of airport and an equivalent amount on maintaining and
developing them. The buildings will be there for decades so we need to ensure that they
will respond to needs of tomorrow. The Airports of the Future will be a response to the
aspects of the future and these are intertwined and interdependent:
Aspect Examples
Environment climate, resources, pollution, noise
Technology communications, users interfaces, intelligent buildings, materials
Future Society global politics, (de)regulation, security, tax, welfare, culture
Future Business globalisation, supply chains, retail, money, employment patterns
Future Passengers demographics, lifestyles, expectations
Future Aviation alliances, aircraft developments, market segmentation, congestion
WHY LOOK AT FLEXIBILITY?
Can we Predict the Future? No. There are plenty of well known examples of famous
people getting it wrong. But we don't have to be able to predict a single future to be doing
useful work: we can design for flexibility.
In some areas we can have more certainty than in others. In the more clear cut cases
we can plan in greater detail with higher confidence; in others we should aim to build in
flexibility. But which are which? And what sort of flexibility is needed where?
If we can't predict the future we should be aiming for flexibility, so that we will be
better poised to cope with future events. If it is done systematically, being prepared for
likely eventualities can give critical business advantages and save large amounts of
money: building in flexibility doesn't always mean adding cost.
LEAN AND FLEXIBLE
A significant body of literature now exists to describe lean production methods as well as
lean construction theory and applications. The weight of this knowledge lies generally
within the construction phase of projects and there is less case history of its application to
design. With certain and known requirements the principles to achieve lean design and
construction are understood. However this may not lead to a flexible solution. Many
development projects fail to meet the expectations of the end users.
Such project failures can be classified into one of five basic types :
1) The solution fails to meet the business requirements for which it was developed. The
solution is either abandoned or expensive adaptive maintenance is undertaken.
2) There are performance shortcomings in the solution, which make it inadequate for the
users needs. Again, it is either abandoned or amended incurring extra costs.
3) Errors appear in the developed solution causing unexpected problems. Modifications
have to be applied at extra cost.
4) Users reject the solution, for political reasons, lack of involvement in its development
or lack of commitment to it.
5) The solution is accepted but over time becomes unmaintainable and so passes into
disuse.
Conversely it is also relatively simple to include flexibility either in the design
process through including extra design iterations, or developing multiple designs or
incorporating redundancy or duplication or provision into the built solution. This however
is not lean or efficient. The solution will probably be late and over budget through
numerous changes.
SYSTEMS THINKING
The ability to see the big picture is fundamental to solving the right problem. Systems
thinking (Senge 1990) is needed to understand the complexity and interactions of the
various parts of a whole framework. For Terminal 5 the goal is to create the worlds
most refreshing interchange .This means that we have to balance the short and long term
views. What might be happening in the future depends on how far ahead we are looking.
The aim is to balance on the one hand looking forwards from today's position by
following current trends, with on the other hand stepping back and trying to create a
vision of the future in say 20 years time. This might sound a long way off, but it's less
than half way through the life of Terminal 5.
We can portray this as follows:
Is it any wonder therefore that the traditional starting point for delivery of a matching
brief and business case tends to be the holy grail that is never achieved on large, complex
and long projects? Yet without this matching set the delivery process can not start and be
lean and efficient. We can usefully use function modeling IDEF0 techniques to explore
the principal elements of the system (Figure 2). This will show what information is
required to provide the sufficient and correct flow to start the delivery process and what
information can be given later to give the business flexibility of response.
Figure 2: T5 as a high level IDEF0 Diagram
This high level function diagram can be decomposed two further levels see Figure 3
Figure3: T5 as a system at levels 2 and 3
Create
the Worlds Most
Refreshing Interchange
Project Development
Business Case
T5 operating efficiently
Satisfied Customers
Satisfied Business
BAA BA Suppliers
Site, Public Inquiry, Government Policy, Regulations
Inputs Outputs
Constraints
Resources
Purpose: To delver T5 to meet the evolving Business Case
Formulate Strategy
Produce Information
Produce Product
Operate
Definition
Design
Manufacture
Assemble
Wicked
Problem
Development
Delivery
T5 AS A SYSTEM
Viewing the Worlds Most Refreshing Interchange as a system then:
The Strategy phase needs to deliver a minimum usable subset of Business, Customer
and Operational requirements to enable the Delivery process to start
and
the Delivery process needs to improve flexibility to allow the Business, Customer and
Operations to respond to changes due to technology or market conditions and still be
lean and efficient
ENABLING DELIVERY
WICKED PROBLEMS, RIGHTEOUS SOLUTIONS
This wonderful title comes from a book (DeGrace & Stahl 1990) which describes
methods of solving software development problems. Software development projects also
contain dynamic and behaviourial complexities and it is therefore useful to look at the
tools and methods that have been adopted in that industry and to draw learning for our
own paradigm. To start the Delivery process we need output from the Strategy phase of
the project. This entails solving wicked problems.
The issues and constraints relating to T5 are shown in Figure 4.
Figure 4: Viability Hierarchy
The viability hierarchy represents the dynamic and behaviourial complexities of the
project and as such there is no ideal solution that meets these all of these requirements.
The satisficing solution,(Simon 1957), is the one good enough to satisfy the
stakeholders. A backward pass down the timeline from an operational open date of 2007
dictates when the first outputs from the strategy phase need to be delivered to the Delivery
process. BAA has for the past 5-6 years adopted a long term partnering/framework
approach with its suppliers this has enabled an integrated team of Customer, Operator and
Functionally
Acceptable
Operationally
Acceptable
Affordable &
Fundable
Meets PI &
time
constraints
Integration
with LHR
Easeof
Increasign
Capacity
Opening
Date
Capacity
Customer
Satisfaction
T5Vision
Safety,
Security &
Environment
Punctuality
Operational
Efficiency
Income
Operating
Cost
Capital Cost
Funding
Basis
PI
Compliance
Differentiation
Show the
World's Most
Refreshing
Interchange is
Viable
Viability Criteria
the Delivery team (comprising of the designer, manufacturer and assembler of all the
primary systems) to develop trial solutions using incremental prototyping .
The prototyping of solutions is undertaken within fixed time cycles. The purpose of
each cycle is to make and test something. Customers , operators and the business are an
active part of each cycle. The output of each cycle comes from a prioritised list according
to MoSCoW rules:
Must Have for requirements that are fundamental to the system. These must be
delivered and represent the minimum usable subset of a solution.
Should Have for important requirements that should be there or a short term work
around is available
Could Have but could be omitted this cycle
Want to have but probably wont for this iteration
The os are just for fun.
After some six cycles over fifteen months there is now emerging the fundamental
requirements that enable the Delivery process to start. Time has now run out for further
iterations unless a delay to the opening is acceptable.
IMPROVING THE DELIVERY PROCESS
Solving Wicked problems means that the Strategy phase is now able to provide a usable
set of input information flows to start the Delivery Process. The Delivery process needs to
respond to its part of the system by allowing flexibility and still be lean and efficient .
Flexibility is improved by the adoption of Last Responsible Moment (LRM) for
information transfer. LRM is analogous to Just In Time (JIT) for construction but can be
universally applied to milestones for transfer of any type of information or product. Lean
Thinking is applied to value streams by improving flow of information or product through
the system and removing waste. Essential but non value adding support processes such as
resourcing and approvals are run concurrently adopting a smile and wave support
approach.
PROJECT PROCESSES RE-EXAMINED
TRADITIONAL PROJECT PROCESS
The standard BAA Project Process is illustrated in figure 5.
Figure 5: The BAA Project Process
It is similar to numerous other project processes. It was developed and works well for
Concept Design Co-ordinated Design
Gate
Earthworks
Structures
CCTV
Gate
Production Drawings
Gate
Construction
Earthworks
Structures
CCTV
Earthworks
Structures
CCTV
Earthworks
Structures
CCTV
medium sized projects with a fixed brief at the start of Concept Design. There are a series
of gates at which the solution is considered by the stakeholders and approval given to
proceed with the next stage.
It has a number of disadvantages for a project such as T5:
It is a batch and queue process which trades efficiency in the delivery process for
security to the business.
It is inflexible and does not work well if the business case and customer requirements
evolve after the start of Concept Design.
It does not inherently cater for fast moving technologies where manufacture and
assembly and therefore design have no need to start early and considerable advantages
may be obtained by delaying design to ensure that the most up-to-date technology is
incorporated.
The high level map is based on traditional construction stages and other disciplines
cannot satisfactorily map their own processes to it.
LAST RESPONSIBLE MOMENT PROJECT PROCESS
The Last Responsible Moment (LRM) is defined as the latest moment for starting an
activity without compromising cost or programme whilst maintaining maximum
flexibility for the Business. The LRMs identify key programme drivers such as:
information flows between systems (such as structural grid from building concept to
structural frame)
technical decisions (choice between pre-stressed or conventional aircraft pavements)
business decisions (trading between retail space and check-in space)
The LRM dates are established by working backwards from end dates for each major
system of the project. These dates are not the last possible dates but change later would
have cost or programme implications. The principles of the LRM Project Delivery
Process is shown in Figure 6
Figure 6: LRM Delivery Project Process
LRM
Manufacture & Assembly
LRM
Design
LRM
Definition
LRM
Develop
Requirements &
planning
Earthworks
LRM LRM LRM LRM
SubStructure
LRM LRM LRM
CCTV
Principles
The key principles for the process are:
Identify key systems for the project (Earthworks, Aircraft Pavements, Substructures,
Building Services, Envelope etc.)
Process map the value streams of systems.
LRMs are milestones for information transfer
Project programming and monitoring is based on the LRMs of systems and the flows
between systems.
The project management system is driven by information flows
INFORMATION DRIVEN PROJECT MANAGEMENT SYSTEM
All activities require information to be processed. Information is required before activities
can be started and each activity produces information once it is completed. Understanding
the flow of information is therefore fundamental to planning and doing work. If we can
track and control information rather than monitor activities we can ensure that the right
people have the right information at the right time and when this does not happen it
pinpoints the source of the difficulty.
THE HIGH LEVEL PROCESS
Figure 7: Information Flows and Black Boxes
The high level process represents the conversion of inputs to outputs by adding value
(Figure 7). Outside of those undertaking the process or activity we do not need to
understand the conversion process other than to ensure that it is as efficient as possible.
Inputs are those pulled by the activity and are milestones. Similarly the outputs are those
required as inputs by following activities within the system being mapped or in other
systems.
MAPPING THE T5 SYSTEMS
Mapping is carried out in at system level with, as far as possible, the Business, Customer,
Operator and Delivery team being represented. The sessions are facilitated and
concentrate on identifying the inputs and outputs working backwards from the Operate
phase back through to Definition (Figure 3). Some 30 systems need to be mapped to
establish the high level LRM programme. Time is allocated to the activities such that
programmes can be created. The process maps are drawn into VISIO and then
automatically transformed into programmes using an internally developed Visual Basic
I
N
P
U
T
S
O
U
T
P
U
T
S
Activity or Process
Time & Resource
programme. Figures 8 & 9 illustrates part of the process map for substructures and the
resulting programme.
Figure 8: Part of Substructures process map
Figure 9: Corresponding part of the Substructures Programme
ID
Scheme
design layouts/
sections/loads
SB3505
ID
Interim details
of pile testing
SB3510
ID
Location of
CTB expansion
joints SB3515
ID
Station layout
& alignment
2nd iteration
SB3575
ID
Baggage cycle
3 SB3580
Detailed CTB
substructures
design
1
SB2705
ID
1st pass
detailed design
for costing
SB5505
Layout, cost &
programme
review &
approval
1
SB2715
Develop
construction
methodology
& phasing 1
SB2710
ID
Programme/
cost restraints
for
construction
SB3520
ID
Construction
methodology &
phasing
SB5510
Develop
technical
requirements
1
SB2890
ID
System
requirements
SB3585
ID
Approval
SB5515
ID
Technical
requirements
SB5545
Complete
detailed CTB
substructures
design 1
SB2720
ID
Complete CTB
Substructures
design SB1005
ID
Final update of
layout,
sections &
details SB3525
ID
Final CTB
Loads SB3530
Activity
ID
Activity
Description
Substructures
CTB & Station Boxes
Detail Design
SB3505 Input - Scheme design layouts/sections/loa
SB3510 Input - Interim details of pile testi
SB3515 Input - Location of CTB expansion join
SB3575 Input - Station layout & alignment 2nd iterat
SB3580 Input - Baggage cycle
SB2705 Detailed CTB substructures desi
SB5505 Output - 1st pass detailed design for cost
SB3520 Input - Programme/cost restraints for con
SB2710 Develop Construction methodology & phas
SB5510 Output - Construction methodology & phas
SB3585 Input - System requiremen
SB2890 Develop technical requiremen
SB5545 Output - Technical requiremen
SB2715 Layout, cost & programme review & approv
SB5515 Output - Approva
SB3525 Input - Final update of layout,sections,deta
SB3530 Input - Final CTB load
SB2720 Complete detailed CTB substructures des
SB1005 Complete CTB detailed substructures des
2000 2001
OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT N
Primavera Systems, Inc.
Start Date 21JUN99
Finish Date 28MAR04
Data Date 27MAR00
Run Date 13APR00 14:18
Late Bar
Progress Bar
Critical Activity
LRM3 - SB01
Substructures
Design Programme
Sheet 1 of 9
The main apparent difference from a conventional programme is that all the inputs
and outputs are milestones and these are used to monitor progress.
WHAT INFORMATION?
The process sessions need to be structured to ensure the correct information is captured:
What are the critical external information flows with other systems(including the
Functional Requirements)?
What are the critical internal information flows that affect the timing of the external
flows?
What are the critical internal information flows that define the Last Responsible
Moment for the start of the next sub-process?
Is the level of detail sufficient to model the critical internal and external flows?

Added value
Dealing with information flows rather than activities gives further advantages. Inputs and
outputs (datasets) are ideally suited to document management systems with control over
versioning and transfer of information and workflow control. With each piece of data
additional information can be attached such as:
Approvals and reviews undertaken
Risk and opportunity (does the data contain an assumption, what are the transfered
risks?)
Interface control information ( what activity provides the information and what
activities receive the information)
The future worth of these datasets is greatly enhanced and facilitates knowledge
retrieval.
SO FAR -SO GOOD
The understanding and the adoption of the principle of LRM and information flows has
been pivotal in allowing the project to move forward. The Business now has an
understanding that decisions need to be made by certain dates to avoid waste and meet
operational timings. The delivery team acknowledges that a comprehensive brief is not
available at the start of the project. Systems thinking has allowed us to see the wood and
the trees. However, much needs to be done and we have only just taken our first few steps
down a long road. We are currently exploring methods of optimising the information
flows which is currently being done using experienced judgement. The use of Design
Structure Matrix (DSM) techniques will be a useful tool. So far we have concentrated the
effort on the strategy and design phases and much process mapping work needs to be
done for the Manufacturing and Assembly phases. We hope to report back in the future.
ACKNOWLEDGEMENTS
Parts of the futures text come from the BAA intranet web site Airports of the Future -a
project managed by Laurence Matthews. Parts of the description of Complexity come
from an unpublished paper by David Hancock who is part of the T5 team.
REFERENCES
DeGrace & Stahl 1990. Wicked Problems, Righteous Solutions-A Catalogue of Modern
Software Engineering Paradigms, Yourdon Press, PTR Prentice Hall Building,
Englewood Cliffs, NJ 07632, USA
Simon, H.(1957). Models of Man. Wiley, New York
Rittel and Webber 1973. Dilemmas in a General Theory of Planning ,pp155-169, Policy
Sciences, Vol 4,Elsevier Scientific Publishing Company,Inc
Senge 1990. The Fifth Discipline - The Art and Practice of the Learning Organisation
Century Business Publications.