Revision Session: Suggested Answers

Module: Marketing Research- June, 2011 Time: Weeks 16 -17

Day 1: Case Question 1: Theme Park (Jun 2009 - M 10)

June 2009 Examiner's Solution

Case 1 Answer Base 1. Explain to the managers of the theme park the value they might derive from multivariate regression analysis of the data. (15 marks) The relevant module is Module 10. Multivariate regression analysis will help the managers of the theme park to identify what factors are important to customers during their visit to the theme park and the relative contribution of these key factors to the development of customer satisfaction. So, for example, multivariate regression analysis can show whether queue management actually affects customer satisfaction and to what extent compared with other variables, e.g. staff friendliness. Multivariate regression analysis allows the development of a predictive equation which shows what changes in overall customer satisfaction may be anticipated as a result of, for example, initiatives to reduce queue waiting time. This will allow the theme park managers to gauge future investment priorities, if development of increased levels of customer satisfaction is a key business objective. The use of interval scales in the market research study allows the use of the multiple regression procedure for the analysis of the theme park market research data. Multiple regression analysis finds the weighted linear combination of independent variables, e.g. queue management, safety, etc., that has the maximum correlation with the dependent variable, which is overall customer satisfaction. In undertaking this analysis, clearly the value of information produced is very much dependent on inclusion of the independent variables in the market research questionnaire. A good student answer will note that if there is doubt that the set of independent variables researched is not fully inclusive in terms of what is important in delivering customer satisfaction in a theme park environment, it may justify qualitative research to identify what these key variables are. If other variables are at work, then the quantitative research must be re-commissioned.
1

2. Are these data helpful in determining investment priorities to increase customer satisfaction levels in future? Explain your reasoning. (15 marks) The regression function takes the form Y = a + b1 X1 + b2 X2 + + b n X n The computer output suggests the following equation for customer satisfaction: Customer satisfaction = 1.0 + (0.6 Queue management) + (0.3 Staff friendliness) (0.3 Security procedures) So, for example, for every scale point that queue management increases, holding other variables constant, customer satisfaction levels increase by 0.6 units. A good student answer will note that, interestingly, for every scale point increase in security procedures, customer satisfaction decreases by 0.3 units. To understand whether these variables are statistically significant, it is necessary to test the significance of the regression coefficients. The individual values of the t-statistic for each of the variables can be compared with the critical t-statistic at the 5% level of significance, i.e. 1.96. Since the calculated t-statistic for queue management (2.73) and staff friendliness (2.50) are greater than the critical t-value of 1.96 at the 5% level of significance, it is concluded that both queue management and staff friendliness are statistically related to customer satisfaction. In considering future investment priorities, improved queue management is likely to have more impact on customer satisfaction than programmes to increase staff friendliness. This of course only holds in the current equation. The values of the coefficients will all change when safety is excluded. A good student answer will note that security procedures are observed to have a negative effect on customer satisfaction, but that the findings are not statistically significant at the 5% level and, therefore, security procedures should be excluded from the regression equation. The regression analysis should be run again and the variables tested for statistical significance. As noted above only queue management, staff friendliness and security procedures have been considered in the multiple regression analysis. If it is the case that other unidentified variables affect customer satisfaction (and the adjusted R square would tend to support this thinking as it suggests that only 40% of the variation in customer satisfaction is explained by queue management, staff friendliness and security procedures), then the theme park managers should investigate what these other variables might be and their relative impact on customer satisfaction levels.

To do so the theme park managers should undertake qualitative research (to identify the key variables) and then commission a further quantitative study to provide ratings across all variables together with measures of customer satisfaction.

June 2009 Selected Student Answer

Case 1 Solution: Theme Park 1. Explain to the managers of the theme park the value they might derive from multivariate regression analysis of the data. Multivariate regression analysis (MRA) is a statistical method of finding whether there is a relationship between independent and dependent variables. This is the equation for multivariate regression: Y = b0 + b1 X1 + b2 X2 + b3 X3 + + b n X n In this case, the MRA will find out the relationship between customer satisfaction with a different set of independent variables, such as queue management, staff friendliness and security procedures. If we find out, for example, that there is a relationship between customer satisfaction and queue management and/or staff friendliness then the manager might have to consider improving the queue management and staff friendliness, but if we find out there is no relationship between customer satisfaction with security procedures then we might consider trying to establish whether there are other independent variables which might affect or have a relationship with customer satisfaction. Multivariate regression analysis helps to analyse the information and it is one of the quantitative research technique. Also there is a need to explain the regression correlation (r) which 1 < r < +1.
o o o

r = 0 This means there is no relationship between x and y r 1 This means there is a relationship between x and y which means if x increases by 1 then y will increase by 1. r 1 This means there is a relationship between x and y which means if x increase by 1 then y will decrease by 1.

Suppose regression analysis, with customer satisfaction as the dependent variable, produced the following output: Aspects operation Staff friendliness of Coefficient Standard error 0.22 0.12 0.27 t-statistic 2.73 2.50 1.11
3

Queue management 0.6 0.3 Security procedures 0.3

Constant Adjusted R square

1.0 0.4

2. Are these data helpful in determining investment priorities to increase customer satisfaction levels in future? Explain your reasoning. In analysing the variable by using regression coefficient analysis we will first set:
o o

H0 = 0 (No relationship between x and y) In this case customer satisfaction and either queue management/staff friendliness/security procedures HA 0 (relationship between x and y)

As % confidence level = 1.96 Degree of freedom = 17.1 t-statistic = 2.73 (for Queue management, 2.50 for Staff Friendliness, and 1.11 for Security Procedures.

If we are using these data to analyse the regression coefficient then we can say that queue management and staff friendliness have a relationship with customer satisfaction. It means that if queue management is improved and staff are more friendly then there will be increase in customer satisfaction. However, to determine an investment we might have to consider other factors or undertake other types of research because it may give information which will be more accurate for the manager to make a decision. First we might consider sample size because a bigger sample gives more accuracy. Also we need to consider whether to use probability or non-probability sampling. In this case we might use stratified sampling. Moreover, we have to design a proper questionnaire which includes the preliminary considerations, age, sex, status, social and economy. Also it have to provide closed-end and open-end question which will give more opportunity to understand feelings and attitudes. Then we have to conduct a survey which is one method of quantitative research but before that we might have to provide a qualitative research in which we might set up a focus group or in-depth interview to understand feelings and attitudes, and, then conduct the quantitative research. The survey can be postal survey, personal survey/interview or telephone survey. In this case, we might decide to use postal survey and personal survey.
4

Day 2:

Case Question 1: Electronic Waste (December 2007 M10)

December 2007 Examiner's Solution

5

Case 1 Answer Base 1. What does the market research tell us about the effectiveness of the Amnesty? (15 marks) The market research gives us only limited insight into the effectiveness of the Amnesty because:
o

the findings of the research study are not representative of the views, attitudes and behaviour of the wider population of companies located in and around Greenville; the marketing research undertaken by Marson Ltd was conducted amongst only those companies taking part in the Amnesty event. the views of non-participants may be quite different to those of participants; it may well be worthwhile to undertake a wider study to ascertain these.

The validity of the findings may be affected by the sampling method used. Sampling error may have arisen from, for example:
o

o o

The way the sample was drawn (for example, were there strict rules to interview every nth visitor, or did interviewers have freedom to choose their interviewees?) The layout of the site and the location of interviewers (for example, was there more than one entry point to Marsons site or site office? Did large trucks use a separate access point? Were all companies visiting the site on the day equally likely to be interviewed or not? Were interviews evenly spread over the day or did interviewers simply do their assigned quota and leave?) Was the individual interviewed the right person to speak to about such matters? There is a big difference between the views of the van driver and those of top management! The number and treatment of spoiled interviews. The level of supervision of the interviewers and whether attempts were made to back-check a sample of the interviews.

In addition, the research undertaken on the Amnesty day may have been subject to nonsampling errors which may also have affected the validity of the research, for example:
o

The market research was conducted by face-to-face interview but no information is provided on whether these were undertaken by professionally trained interviewers were questions asked in a leading way? Were responses recorded correctly? Did any tabulation errors arise during data input and analysis?

A good student answer will note that the focus of the market research is only on those taking part in the Amnesty and, therefore, the key findings of the research can only be considered in this context. The market research, therefore, tells us little about the effectiveness of the Amnesty generally but, assuming satisfactory resolution of the concerns relating to the potential for sampling and non-sampling error in the implementation of the research, it is possible to draw some statistically sound conclusions from the research. The richness of the analysis will depend largely on the questions asked
6

and careful profiling of the companies participating in the research (details of which we do not have). Of interest, for example, might be to analyse by type and size of company: type of equipment handed in average amount of equipment handed in distance travelled to the recycling site difference in awareness of recycling possibilities willingness to pay for disposal/recycling proportion of companies who clean own PC systems before hand-in proportion of companies seeking security reassurances 2. Is there a statistically significant association between size of company and awareness that some electronic equipment can be recycled? Explain the implications of your answer.
o o o o o o o

(15 marks) A chi-square test can be used to test whether a statistically significant association exists between size of company and awareness that some electrical equipment can be recycled. The chi-square statistic is calculated by the formula:

where O i is the observed value and E i is the expected value, assuming in this case that there is no difference between the backgrounds of respondents. For a particular cell, the expected value is calculated as:

Awareness that some electronic equipment can be recycled Company size Large (O = observed) (E = expected) Small (O = observed) (E = expected) Total Not aware 862 759 1018 1122 1880 Aware 288 391 682 578 970 2850 1700 Total 1150

We hypothesise that there is no relationship between awareness that some electronic equipment can be recycled and company size. The chi-square statistic at 1 degree of freedom at the 5% level is 3.84 and because 69.46 > 3.84, the difference noted in the sample is statistically significant. The research findings indicate that 59% of small companies, and 74% of large companies, are not aware that some electronic equipment can be recycled. The chi-square test reveals that the difference in awareness is statistically significant. A good student answer will note the importance of this finding in terms of the methods used to communicate recycling opportunities. It may well be the case that the involvement of more people in waste management issues in larger companies makes targeting of messages about recycling opportunities more difficult and that a multi media strategy is required to reach all of the individuals concerned. On the other hand, this research finding may reflect another problem in the way the sample was drawn for the survey. In large companies waste management is likely to be a delegated responsibility; the individual visiting the Marson site and interviewed on Amnesty day may be well down the decision-making hierarchy and, therefore, not knowledgeable about such matters, i.e. their views cannot be held to be representative of those held by senior management of the company.

December 2007 Selected Student Answer

Case 1 Solution: Electronic Waste 1. What does the market research tell us about the effectiveness of the Amnesty? In conducting any marketing research it is always important to clearly define what the objective of the research is. In the case in question, the aim of the amnesty was to publicise that irresponsible disposal of electronic waste is bad for the environment and that a lot of electronic equipment, particularly personal computers, can be recycled and have second life market potential. It was felt the electronic recycling companies could do much more to raise awareness of proper disposal of electronic waste. The method employed to conduct the research on the day of the amnesty was based on face-to-face interviews. The market research informs us that all those interviewed felt the amnesty was a good idea. However it must be noted that these were the persons who attended and a good result of the research would have considered a population frame that included other persons in the community who may have had a different view or a suggestion not thought about by the recycling company. A random sample was taken. However we are not sure of the questions on the structured research questionnaire that were asked. The research appeared to be simply to confirm in
8

the minds of the researcher that the amnesty was a good undertaking. Marketing research must be undertaken to guide the decision of the manager and not simply to confirm what one wants to do. The sample was taken from persons who handed in unwanted electrical or electronic equipment on the day and this may have been a way of helping those persons to get rid of obsolete equipment and not necessarily that their intentions supported the objective of the researcher. However the fact that 2850 companies handed in unwanted equipment is a success for the company (Marson Ltd). But there is no knowledge of the population of persons (companies) with unwanted equipment on hand. The fact that 84% of the companies taking part were not aware that recycling was available informs the researcher that there is a need for more advertising and possibly promotional opportunities to ensure that the knowledge is more widespread. 55% of the companies felt that Marson Ltd could charge companies a small fee for disposal and recycling raises an opportunity for a business venture for Marson Ltd and the research may be considered effective to some level. However further research will be necessary. A combination of focus group and a more structured survey of companies, possibly by mail (since companies are involved and these are busy people), or by mixing with telephone interviews. While 55% is a notable number of persons, we must remember that the interview was conducted on the day of the amnesty and only the persons who were interested enough attended; and yet only 55% of those who attended were of the view. 68% of companies sought assurance of hard drives being checked and cleaned before recycling may be considered to give Marson Ltd extra work to be done and of course at a cost. This would have to be evaluated further by the company to see whether that consideration is one that they had thought about. In conclusion, it is difficult to say whether the research was effective. As noted before, marketing research must be undertaken in a structured manner and objectives must be clearly defined to identify the information objective. Parameters must also be set before research is undertaken. This is in respect to budget and related issues to the success of the research. The entire process must be clearly planned including the sampling method and the sample design. The population units and boundaries were previously touched on. The analysis of the report is also important bearing in mind the objectives that were outlined. The market research essentially does not tell us much about the amnestys effectiveness because there are several unknowns in the equation. 2. Is there a statistically significant association between size of company and awareness that some electronic equipment can be recycled? Explain the implications of your answer. A chi-squared analysis of the data could be conducted to determine whether there is an association between the two data groups. Based on the findings, a hypothesis can be drawn. This is usually represented as follows: HO HA association no association

An expected frequency can be calculated using the following formulae where:

The chi-square statistic can then be calculated using the following formulae where:

The chi-critical statistic is then calculated as follows df (rows 1) (columns 1) where df refers to degrees of freedom. According to the information provided the chi-critical statistic at 1 degree of freedom at the five percent level is 3.84. Company size Large (O = observed) (E = expected) Small (O = observed) (E = expected) Total Not aware 862 758.5 1018 1121.4 1880 Aware 288 391.4 682 578.59 970 2850 1700 Total 1150

Using the formulae for expected frequency:

Calculations

Calculations

10

Chi statistic at one degree of freedom is 3.84. According to the calculations with chi-squared statistic at 69.44 being higher than chi critical at 3.84, a HA hypothesis can be drawn where there seems to be an association between company size and awareness. There needs to be greater communication between WEEE and companies. Suggest they organise visits to all companies to inform of the standards and efforts to better ensure compliance.

Day 3: Dec 2010 exam questions

Case Question 1: Fruity Juice Bubbles

December 2010 Examiner's Solution

Case 1 Answer

11

1. Advise Pete Martin on how he might use focus groups to help develop a marketing strategy for Fruity Juice Bubbles. (20 marks) The relevant module is Module 8. It is appropriate to first consider what focus groups are and the general procedures used in running focus groups, including the role of the moderator in facilitating the discussion As far as process is concerned, the focus group method relies on self-reports from individuals in a group situation which are moderated by a professional market researcher. The group moderator uses a semi-structured, and sometimes unstructured, approach to discussion of a subject a method designed to reveal respondents full flow of thought in relation to the subject matter. The intention in focus group interviewing is not just to ascertain what peoples overall views and opinions are, but also to understand why people think, feel and behave in the way they do. The different types of focus group are: exploratory, clinical and experiencing. As far as Fruity Juice Bubbles is concerned, and in line with the aim of the research which is to help guide the development of marketing strategy, developing a better understanding of the target customer is vital. The exploratory or experiencing format looks to be most appropriate. The focus groups will provide an opportunity for Pete to develop his understanding of target audience motivations and purchasing patterns (how often to buy, where to buy, purchase occasions, etc.), the product attributes considered to be most important to the target audience (determinant attributes), the competitive set and the relative positions held by different carbonated soft drinks, the strength of the various positions held by competitive brands in this category, and potentially attractive positions for Fruity Juice Bubbles. 2. Prepare a focus group discussion guide for the research. (10 marks) The discussion guide will include a welcome and introduction to the subject to be discussed. The facilitator should set respondents at their ease and explain at the outset that there are no correct opinions, and that each group members comments and opinions are valid. The guide to the focus group discussion for Fruity Juice Bubbles will reflect the aim of the research, which is to help guide the development of marketing strategy. Understanding consumer behaviour in the carbonated soft drinks category is key, as is the development of a clear picture of the perceived positions held by competitive brands and potential positions which offer marketing opportunity to Fruity Juice Bubbles. An example of a discussion group guide for Fruity Juice Bubbles would be: 1. Opening Explain focus groups No correct opinions, only opinions Need to hear from everyone
12

2.

3.

4. 5.

Audiotapes Procedure one person to talk at a time, etc. Any questions Carbonated soft drinks Which carbonated soft drinks are consumers aware of, which have they tried and which do they like? What like/dislike about individual brands? When drink (usage occasions)? Where purchase? Who purchases? Individual item purchase or multi-pack? Value for money perceptions Image of individual brands What are the important and determinant attributes in purchase decisionmaking and reasons? Fruit Juice Bubbles Initial reaction to product concept Perception of key competition Is there a gap in the market? Distinctiveness of concept What like/dislike about concept? Likelihood of trying and perceptions of usage occasion Reaction to intended price premium Tasting opportunity reaction to taste and range of flavours Reaction to proposed size(s) of bottle Where would expect to buy, etc.? Final questions and comments Thank you for your co-operation

December 2010 Selected Student Answer

Case 1 Solution: Fruity Juice Bubbles 1. Advise Pete Martin on how he might use focus groups to help develop a marketing strategy for Fruity Juice Bubbles. Nowadays, when almost each market is highly competitive and rapidly changing, when customers needs and wants become more and more sophisticated, it is becoming more and more difficult for companies to compete. So, understanding customers needs and wants, attitudes, preferences and beliefs in order to match them is essential. To do this marketing research can help. Talking about our case, the company is going to launch a new product, so they have to understand if the product will succeed. The product is carbonated soft drinks. As I understand from this case, the product concept is already developed, so the company has to know if the customers will like it.
13

In order to understand customers attitudes, motivations and preferences qualitative research can be conducted. The most popular and the most appropriate technique of qualitative research in this case is focus groups. Focus groups this is when 712 people are led by a moderator discussing particular topics. There are 3 types of focus groups:
1. Exploratory helpful in generation of new ideas. 2. Clinical used in order to reveal peoples inner attitudes and feelings. 3. Experiencing this allows us to evaluate customers during product use, their

emotions, feelings, likes and dislikes concerning the product. The last type of focus group is what our company needs, as product concept is already developed, people have to buy it and give their judgements. What they like about the product, what they dislike, what they would like to change or add, how likely they will buy this product, how often and in what quantities. Answers on all these questions will help the company to match their potential customers needs ideally, so becoming successful and increasing profits. Focus group discussions can help not only in product evaluation. Pete Martin is developing a marketing strategy and focus groups can help him to develop it in the best way, related to product. Focus groups can discuss such aspects as preferable set of marketing mix 4 Ps. I already mentioned the product, the rest are: Price at what price they are likely to buy the product. Place where it is better to sell the product in order to reach availability and convenience, Promotion what kind of advertising can be attractive, create awareness and goodwill around the product. Additional decisions could be made about the packaging of the product it should be attractive, as well as convenient, and eye-catching. During the discussion, the moderator can have an influence, so it is a position with high responsibility, and the person has to understand this. The moderator has to be well trained, with a quick mind, a good memory on names, socially acceptable and as neutral as possible in order to avoid negative influence and biased results. Once the data of research is collected and analysed, the results can help Pete Martin to develop an appropriate strategy and compete on the market successfully. 2. Prepare a focus group discussion guide for the research. As I mentioned above, the valid results in most cases depend on the moderator, who leads the group discussions. He has to be well-trained, be able to identify if respondents have enough knowledge about the topic, to quieten those who are talkative and encourage to talk those who are quiet. He doesnt have to push on the people, avoid biased words and be as neutral as possible. He has to have a quick mind in order to catch good ideas and lead a discussion in a necessary way. At the start of the discussion he has to warm up the group. Create a warm friendly atmosphere, and make people feel relaxed and safe.

14

He has to identify the level of knowledge of respondents about the topic. Maybe some screening questions should be asked. Then the discussion on the main topic is to follow. In our case, it is a new product, so following questions should be asked: o o o o o o o o Do they like the product? What they dislike about it? What they would like to change in it? At what price they would likely to buy the product? How often and in what quantities? Where it would be convenient to them to buy the product? What kind of advertising can attract them and encourage to buy the product? Will they advise the product to their friends, relatives, etc.?

After the needed information is obtained, the moderator has to give a possibility to ask any last questions. It is necessary to summarise what was said in order to check if everythings understood correctly. Then thank them for participating and warmly finish the session. The better the atmosphere in the group, the better results obtained.

Case Question 2: Business Solutions Ltd.

December 2010 Examiner's Solution

Case 2 Answer 1. Critically explain the threats to experimental validity.
15

(15 marks) An experiment implies some sort of test to allow the effects of independent variables on a dependent variable to be discerned. Dependent variables are the outcomes of interest in this case, visits to the website. The independent variable, i.e. the variable which the researcher has some control over is the telesales initiative. The use of an experiment will help Business Solutions Ltd. to decide whether a telesales initiative will increase visits to the website. In examining whether there is a causal relationship between telesales activity and visits to the website, three types of evidence are important:
o o o

concomitant variation, which influences the extent to which telesales activity and web sites visits occur together in a predictable way time order of occurrence of telesales activity and the effect on website visits absence of other causal factors, i.e. no other factor is affecting website visits

Experimental validity has two components:

o o

Internal validity External validity

Internal validity is the extent to which an experiment controls the effects of all nonmanipulated variables so that any difference in web site visits (the dependent variable) between groups can be regarded as valid effects of the different promotional programmes used (experimental factor). Threats to internal validity include:
o o o o o o

History and maturation Repeated testing Impact of the researchers Mortality of participants Selection errors Regression effects

Each of these effects should be considered in the context of the case situation. For example, in relation to the impact of researchers, the fact that an SME is being interviewed may affect visits to Business Solutions Ltd website. External validity refers to the extent to which experimental effects will generalise to the marketplace. Whereas a laboratory experiment is generally believed to be more internally valid, the field experiment (as used in the case situation) provides more external validity as it is undertaken in a real market situation. As the experiment is being undertaken in the field, it offers more realistic conditions in which to evaluate the effect of the supporting telesales campaign. 2. What type of experimental design do you suggest Business Solutions Ltd should use? Explain your reasoning.
16

(15 marks) The different types of experimental designs are:

o o

Basic (informal) designs: only measure the treatment impact. Statistical (formal) designs: measure treatment impact and also other factors that may have been at work during the experiment.

Basic designs can be described as follows:

o o o o o o o

After only Beforeafter without control Beforeafter with control Afteronly with control Ex-post facto designs: groups chosen only after treatment has been applied Four group six-study design Time series design

Regarding the statistical designs, they are as follows:

o o o o

Completely randomised design: only takes account of one variable. Randomised block design: measures treatment plus one extra variable. Latin square: measures treatment and two extra variables. Factorial designs: takes into consideration the fact that variables may interact with each other.

The best approach for Business Solutions Ltd. may be the use of the basic design, beforeafter with control. By using the design beforeafter with control, the control group should be comparable in make-up to the group who will receive the treatment (tele-sales activity). The major assumptions in the design are that both groups are affected in a similar manner by extraneous factors. For this to be the case, the subjects must be selected at random and the treatment should be randomly assigned. The statement of the design is:
o o

EG: O1 O2 CG: O3 O4

The effect of the treatment may be measured as:

o

(O2 O1) (O4 O3

Therefore, if the following visits occurred: O1 = 45 visits O2 = 68 visits O3 = 34 visits O4 = 42 visits

17

The assumption is that the increase in website visits in the control group (8 visits) was caused by extraneous factors. If this impact is subtracted from the change in website visits in the companies contacted by the telesales team (23 website visits), the resulting 19 website visits can be viewed as the direct impact of the treatment. It may also be helpful to use a statistical design (randomised block design) to take account of other variables that may influence visits to the website apart from telesales activity.

December 2010 Selected Student Answer

Case 2 Solution: Business Solutions Ltd. 1. Critically explain the threats to experimental validity. An experiment involves a test to determine the causal relationship between variables i.e. the dependent and in this case interest in and visits to the website. The independent, in this case invest in telesales and the effect or results. All experiments must have both external and internal validity. Internal validity refers to the extent to which the effect of the dependent (x) was caused by the independent (treatment y): this takes into account the confidence that the effect of x was not caused by other extraneous factors. External validity on the other hand, refers to the extent the findings/results from the experiment can hold for other situations. An experiment conducted in a laboratory lacks external validity, whilst an experiment conducted in the field lacks internal validity. Hence in this case, if the company Business Solutions Ltd were to conduct a field experiment, this experiment would lack internal experimental validity as the results may have been caused by other extraneous factors. Extraneous factors are factors that affect the dependent variable other than the independent i.e. the treatment. Some extraneous factors are: regression effects (the use of extreme values which skew the results), the influence of the interviewer, selection errors, the maturation and history, repeat testing and the mortality rate of respondents. 2. What type of experimental design do you suggest Business Solutions Ltd should use? Explain your reasoning. There are two types of experimental design i.e. (i) basic design (informal) and (ii) statistical design (formal).
18

The basic informal design looks primarily at the effect of the treatment and not extraneous factors, whilst the statistical/formal designs look at both the affect of the treatment in addition to extraneous factors. Examples of basic/informal experimental designs include:
o o o o o o o

After only ( O), Beforeafter without control O1 O2 (O2 O1), Beforeafter with control, EG (experiment group) O1 O2, CG (control group) O3 O4 (O2 O1) (O4 O3), After only with control (EG O1) (O1 O2) (CG O1), and ex-factor post design which is a hybrid of after only design.

The statistical/formal design examples include completely randomised design, randomised design and Latin square design just to name a few. The completely randomised assumes ceteris paribus and applies the treatment to all the elements randomly and then analyses the results. Given the fact that the Business Solutions Ltd wishes to draw a simple random sample and divided into two groups, where both groups are exposed to direct mail brochure, but only one group is exposed to the telesales activity, I would suggest Business Solutions Ltd use a before-after with control group to determine the effects of the treatment, which in this case is the telesales activity. However, it should be born in mind the caused components i.e. concomitant effect, time, order of occurrence and absence of other factors, if this experiment is to hold on both internal and external validity, hence it may be advisable to use triangulation i.e. more than one experimental design, perhaps group design, bearing in mind the cost and time involved, as well as one other statistical design.

Essay Question The Uganda Coffee Development Authority

December 2010 Examiner's Solution

Essay Solution Answer Base 1. Advise The Ugandan Coffee Development Authority on the need for, and design of, a marketing information system to help realise their goals. (20 marks) The relevant module is Module 16.
19

A marketing information system is vital to help The Ugandan Coffee Development Authority to make good decisions and provide sound research and information to their various stakeholders (exporters, farmers, processors, roasters, etc.). A well designed information system enables the efficient handling, organisation and storage of data and may comprise the following sub-systems:
o o o o o

an internal reporting system a marketing intelligence system a market-research system an analytical marketing system.

Useful sources of data will include both internal and external sources: for example, economic trends, political trends, competitor activity, socio-cultural environment and impact on coffee consumption, physical environment, transportation costs, members sales reports, market prices, export data, government sources, etc. 2. What tools do you suggest the Authority use to help forecast market demand for Ugandan coffee? (20 marks) Various forecasting tools may be used to help forecast market demand. In choosing which to use the following considerations should be taken into account:
o o o o

Time horizon Technical sophistication Cost Quality of data that can be used

The basic approach is to make an environmental forecast before making a forecast of sales and profits. Methods to be considered include: asking people questions (surveys of buyer intentions, composite of sales force opinion, expert opinion, analysing past data) and time dependent approaches (classical time-series, moving average, exponential smoothing, and statistical demand analysis).

December 2010 Selected Student Answer

Essay Solution 1. Advise The Ugandan Coffee Development Authority on the need for, and design of, a marketing information system to help realise their goals.

20

The Ugandan Coffee Development Authority has many goals and objectives that they are seeking to achieve. They are striving to excel in their coffee industry and at the same time keep abreast of the dynamics within the environment. A Marketing Information System will assist the Ugandan Coffee Development Authority greatly in achieving their initiatives. A Marking Information System deals with a constant flow of information and coordination of this information to assist decision makers, and to provide the best possible information in an effort to stay abreast in ones industry. A Marketing Information System is made up of four aspects: internal data, marketing intelligence, market research systems and analytical systems. These four aspects produce a sufficient flow of data to decision makers so that they can make the most suitable and efficient decisions. The Ugandan Coffee Development Authority can use each aspect of the Marketing Information System to help them achieve these goals. In terms of internal data this includes things such as information on inbound logistics, outbound logistics, production and operations, customer information and complaints and sales and marketing. Information from all these internal sources can be used to help the Ugandan Coffee Development Authority achieve their objectives. For instance, in an effort to promote, improve and monitor marketing of coffee to optimise foreign exchange, special attention can be paid to sales and marketing information on how sales are doing and the strategies being used by marketing can be assessed. How people respond to different marketing strategies can be assessed and if ineffective it can be tweaked. Information on sales for example can lead the Ugandan Coffee Development Authority to recognise their most profitable segments in their domestic market. They can assess what is attributing to success in these segments and try to transfer the source of success to other segments. They can also look at production and operations information and determine the best, most efficient ways to produce and to provide the best coffee. A very important source of internal information is from customers. Customer complaints, customer enquiries and customer feedback can all be used to develop the product and even give the Ugandan Coffee Development Authority ideas on how to promote coffee as a value added product. It can also give them information on how the domestic market feel about coffee and what can be done to promote their consumption. It can therefore be seen that an abundance of information can be attained internally to help the Ugandan Coffee Development Authority make decisions. Information can also be gained externally. Such information is especially important since the dynamics of the environment are constantly changing. Marketing intelligence and market research systems aid in accumulating helpful information about the industry, about competitors, about consumers (both users of coffee and non-users). Marketing intelligence is the knowledge gained by a firm from everyday interaction with the environment. It can entail information on competitive behaviour, consumer behaviour, sales force attitudes and behaviour, and so on. Having a proper understanding of such information can help the Ugandan Coffee Development Authority understand the environment and industry and thus set the necessary objectives to maintain competitiveness. It can also assist in gaining a competitive advantage over others. Market research systems are another integral part of the Marketing Information System that can greatly assist decision makers in solving problems and setting and achieving objectives. Market research involves collecting, analysing and interpreting data from the environment in an effort to solve problems, meet objectives and understand consumers.
21

Information on answers received from marketing research therefore can help the Ugandan Coffee Development Authority promote domestic coffee consumption. Information from market research can help them segment the market and thus have a more focused approach to satisfying needs. This would inevitably improve the coffee industry. Market research information can also be used in gaining information in international foreign markets. The Ugandan Coffee Development can determine if they are meeting international standards, if the quality is good, how can their coffee be marketed better, and so on. Other external sources of information may include commercial information such as scanners in stores. This can give information of where sales are concentrated, which can help the Ugandan Coffee Development Authority in setting distribution and marketing objectives. Channel distribution information from the sales force can also be attained to help the decision makers in choosing the most suitable channels and selling methods. A final aspect of the Marketing Information System is the analytical systems. These are basically decision support mechanisms that help decision makers in making valuable decisions from the information that they have received throughout the Marketing Information System. It entails forecasting tools, mathematical tools and expert systems. Forecasting tools for example help forecast demand. This can greatly benefit the Ugandan Coffee Development Authority since they can use these forecasts to set appropriate objectives and develop certain aspects of their strategy to meet future demand. The Ugandan Coffee Development Authority can therefore benefit immensely from designing an appropriate Marketing Information System. Special emphasis must be placed on the aspects of the system that give relevant feedback on customers and the coffee industry. All information gained from the different components of the system can be used to help realise goals. 2. What tools do you suggest the Authority use to help forecast market demand for Ugandan coffee? Forecasting demand is an essential aspect of a firm. It allows a firm to recognise the potential of their market and how they can make the necessary arrangements to meet this forecast or even improve this forecast. Sales can be forecasted by a time period e.g. short term, long term, medium term or by location e.g. locally, internationally, or regionally. There are many different methods to use when forecasting. The analytical system of a Marketing Information System deals with the different methods that can be used to forecast sales. Forecasting tools in this system can either be basic or statistical. In terms of basic tools this deals with asking people questions, asking sales forces valid questions, asking experts valid questions and gaining information by looking at past data. Statistically, forecasting can be done by using time series analysis, moving average, extrapolation and statistical demand analysis. The Ugandan Coffee Development Authority can use any one of these methods in assisting them to forecast market demand for their coffee. For instance, they can embark on asking customers about the coffee. This can be done via questionnaires, focus groups, telephone interviews and so on. Finding such information is valid since it allows the Ugandan Coffee Development Authority to
22

recognise how probable an increase in demand is. Asking sales people questions is also very relevant since they are the ones who are in direct contact with the market whose sales you are trying to forecast. Sales force can give information on how and where demand would most probably come from. The Ugandan Coffee Development Authority can also seek help and forecast using estimates from experts. Such experts may be ones from similar industries who may be familiar with coffee drinkers and their habits, information on past demand can also be used to forecast what kind of demand is to come. The Ugandan Coffee Development Authority can also take a statistical approach and use methods such as time series analysis, moving average, extrapolation and statistical demand. Time series analysis for instance is based on the assumption that sales change due to different influences over time. In time series analysis sales can either be seasonal, erratic, follow a trend or rise and fall at different times. Using this method may not be very appropriate for the Ugandan Coffee Development Authority since they are more concerned about forecasting demand of a potential market rather than really focusing on the demand of the current market. A good approach to use statistically is the statistical demand analysis which deals with looking at all the external influences that may affect demand. It looks at things such as personal income, advertising and so on to determine future demand. Ugandan Coffee Development Authority can use this since they can ascertain how these influences affect demand and use it to project future demand. They may also manipulate these influences to further increase demand. The statistical demand analysis balances the equation where S = a0 + b1P + b2A In conclusion Ugandan Coffee Development Authority can collaborate both statistical ways and non-statistical methods of forecasting demand.

Day 4: Case Question 1: Software House (Module 4 June 2008)

June 2008 Examiner's Solution

Case 1 Answer Base 1. How confident should Steven be in the results of his own market research? (8 marks) Steven has undertaken market research already, based on a sample of 80 customers, and the results appear to be favourable to the introduction of interactive voice response technology, with 70% of customers in support of the idea.
23

The question is how confident can Steven be in these results. The Confidence Interval Approach to sample size can assist the evaluation of whether these results are satisfactory as a basis for such investment decisions.

Therefore, we can be 95% certain that the true population proportion of those in favour of the introduction of telephone based voice response is between 60% and 80%. 2. If Steven were to undertake additional research and wished the percentage in favour to be accurate to within 5%, what size of sample do you recommend? (8 marks) One option available to Steven is to conduct further research, increasing the sample size. To be within 5% of 70% Steven would require a sample size as follows:

3. What other methods might Steven use to determine sample size? (6 marks) Other methods of deciding on sample size include:
o o o

intuition cost limitation industry standards

Each of these should be discussed with reference to section 4.6.1.1, 4.6.1.3 and 4.6.1.4. 4. Do you think Steven should press on with the installation of the voice response service or should he conduct more market research? Explain your reasoning. (8 marks) In deciding whether to press on with the installation of interactive voice response services or conduct more market research, Steven should consider the confidence level attached to his existing market research findings. The key issue he has to consider is the extent to
24

which his comfort zone would be enhanced if the confidence interval were narrowed by conducting further market research amongst his customers. Increasing the sample size to 323 would produce a 5% confidence interval which Steven may feel is enough for him to feel confident in making the investment. The cost of undertaking additional market research, i.e. conducting telephone interviews, is not likely to be prohibitive. Steven should take into account the cost of research and the value of the investment decision being undertaken. A good student answer will note that Steven should keep in mind that the survey referred to in the article is a year old and that, in the past year alone, many changes are likely to have taken place in consumer reaction to modern technology. This in itself may explain the differences between the findings of the two surveys.

June 2008 Selected Student Answer

Case 1 Solution: Software House 1. How confident should Steven be in the results of his own market research? Market research helps the supplier to allay uncertainty inherent in all marketing decisions. Relating to the case study, Steven is trying to assess if he should go ahead and purchase a new interactive voice response service (bearing in mind the benefits that his may bring to his company) or decide against purchasing due to articles that mention the frustration of consumers with automated telephone systems. The first consideration in evaluating his research methods is sampling. The research used a probability sampling (where elements in the sample have a known chance of being selected) and therefore it is possible to estimate sampling error. In the case, a very small sample was used (only 80 customers out of his total customer base=2500) and this may influence the accuracy of results or quality of the data. To overcome this flaw, Steven should increase his sampling size because larger samples imply smaller sampling error. It is worth mentioning that the total survey error comprises sampling and non-sampling error. Sampling error refers to the sample not being a perfect representation of the target population and non sampling error can be either non response error (not all in the sample do in fact respond) and response error (respondents giving inaccurate answers). Therefore, although Steven has used a probability sampling and error can be estimated, the small sample size may have affected the quality of the research and hence misled the results. It may well be that the 70% represents the opinion of those who have strong feelings about the subject, or those who feel comfortable with the technology (interactive voice response service) proposed. We should also take into consideration that telephone interviews may cause annoyance to the consumer if not properly conducted and if that is
25

the case maybe all those 70% in favour of the new system just replied positively because they wanted to finish the interview as soon as possible. In short, to produce more accurate results Steven should increase the sample size. Lets calculate the Stevens study accuracy:

If Steven would be satisfied with 10% accuracy which translates to the lowest 60% and highest 80% of satisfaction then he could go ahead. 2. If Steven were to undertake additional research and wished the percentage in favour to be accurate to within 5%, what size of sample do you recommend? To calculate the size sample we can use the level of accuracy formula. It is as follows: L = Z Sp

Notes: L = level of Sp = standard deviation Z = value given for a determined level of confidence precision for proportion

Therefore the sample size at 5% level of confidence should be n = 324. 3. What other methods might Steven use to determine sample size? The methods used to calculate sample size are:
o o o o o

Intuition Statistical precision Cost limitations Industry standards Other factors

26

Intuition is based on informed intuition but its not the most satisfactorily method. Its used in industrial marketing research more often. Statistical precision take into consideration that larger samples are more accurate and that data that flows from it have more quality. It also takes account of the level of precision required ( L = Z S x or L = Z S p) and the confidence interval (the interval within which a population means or proportion have a stated probability of lying). Therefore, this method ticks all the boxes and provides a more accurate result. Cost limitations only consider the available budget despite all the other factors (e.g. level of precision). Therefore, it should only be used when there is no other possibility. Industry standards refers to the rules of thumb developed by experts in the industry. Other factors concern any other variable that may affect/influence how to calculate the sample size. 4. Do you think Steven should press on with the installation of the voice response service or should he conduct more market research? Explain your reasoning. Steven should undertake more research before going ahead with the installation of the new system. His first interview used a sample too small and therefore its results may be misleading. Considering the benefits that the new system may bring, but also the costs that the firm will incur (and bearing in mind all the articles that show a negative feeling from customers towards automated telephone systems), Steven should undertake more research but this time increasing the sample size to decrease sampling error and provide more accurate results. If after undertaking research using the same method telephone survey with a larger sample of respondents he still was not convinced about which decision to make, he may consider the use of qualitative research to complement his quantitative research. Through focus groups they may assess attitudes and opinions on the subject (automated telephone systems) and what underlies these feelings. This more detailed research could provide valuable insights on how the firm may contribute (for those responsible for the development and design of those systems) to a positive customer experience, thus changing the perceptions of those who feel frustrated and disappointed when dealing with this system (maybe the hatred comes from a lack of understanding or ability to use the system and may be easily resolved).

27

Case Question 2: Blood Test Equipment (Module 4 and 8 - June 2006)

June 2006 Examiner's Solution

Case 2 Answer Base Advise the manufacturer of the new home kit on what kind of marketing research they should undertake to develop an understanding of the views and opinions of young people towards heart disease and the use of home testing kits to assess personal risk of contracting the disease. (Total 30 marks) The terms of reference of the marketing research would seem to indicate a need for data of a qualitative nature in relation to the attitudes of young people towards heart disease and use of home test kits to assess personal risk. Clearly, the issues are of a sensitive nature and attitudes may be influenced by personal circumstances, e.g. death rates/disease rates amongst family and friends. Because of the emotional, and potentially distressing, background to the research it is likely that in-depth personal interviewing would be the most appropriate qualitative research technique (other suitable alternative methods may also be discussed).
28

The full range of advantages of in-depth interviewing should be discussed together with their drawbacks, including the fact that they are expensive and time consuming and will require considerable preparatory work in terms of sample recruitment and selection. The need to involve medical practitioners should be considered. Once an understanding of attitudes and opinions has been achieved, the qualitative research may be followed up by a quantitative study to assess the extent to which various views and opinions are held amongst the wider population. The merits of different survey methods should be assessed in the context of the case and some attempt should be made to explain desirable sampling techniques and the range of issues which would lend themselves to quantitative examination.

une 2006 Selected Student Answer

Case 2 Solution: Blood Test Equipment Advise the manufacturer of the new home kit on what kind of marketing research they should undertake to develop an understanding of the views and opinions of young people towards heart disease and the use of home testing kits to assess personal risk of contracting the disease. This is quite a sensitive and serious issue that involves young people. To be able to get the best out of research, I would recommend the use of both Quantitative and Qualitative research methods. Quantitative research is research that is based on numerical or statistical figures. This can be undertaken at the initial phase of the research to ascertain the statistics of how many young people actually have attacks, what treatments actually exist and what specialists exist to deal with this situation. This data will be obtained by the use of secondary data. This can be gathered from health institutions and government bodies such as the Ministry of Health. This will also assist us to define the population the group that fall into our research study for our following qualitative research. The qualitative market research I would recommend to be undertaken is personal interviews and telephone interviews. 1. Personal Interviews Advantages
o o o

Allow interviewer to observe reaction to questions. Enables interviewer to ask questions in a systematic order. Interviewer can steer off sensitive questions depending on reaction of interviewee.
29

Disadvantages Interviewer can be biased. Personality of interviewer can influence answers. 2. Telephone Interviews
o o

Advantages Eliminates interviewer influence. Interviewee can express self freely with the knowledge that interviewer does not know them personally. o High response rate since most people have impulse to pick up the phone when it rings and especially so when it is to deal with young people. 3. Depth interviews such as focus groups can be used as well to identify feelings and attitudes to heart disease and the home testing kit.
o o

Selecting the Sample The success of the research mostly depends on how large the sample size is, and how representative of the population. Samples may be non-probability or probability. Non-probability Sampling Here the sample is not random, but rather managed for practical reasons. Examples of nonprobability sampling are: 1. Quota sampling; 2. Convenience sampling; 3. Judgement sampling. Probability Sampling Here there is a known chance of selection. This is seen as most appropriate since it avoids bias in the selection process and hence is most representative. Some examples of probability sampling are: 1. Systematic sampling. For example, every 80th number from a telephone book is selected as part of the sample. 2. Random sampling. For example, the names of all the students in a university are written on various sheets of paper and put together in a basket or container. The first 40 names that are picked are used as the sample. 3. Stratified sampling. The sample is stratified into e.g. social classes and a number randomly selected. 4. Cluster sampling. This is as stratified sampling except that in this case, the cluster is more demographic. In the case of the blood test equipment, I would recommend that the probability sampling technique be used for the reasons stated above. The specific method I would recommend is the stratified sampling technique. This will give perspective from different sections of social class. This will affect pricing, promotions message and means, distribution methods and where to actually sell the product.
30

Essay Question: Inter-European (Module 4 Dec 2008)

December 2008 Examiner's Solution

Essay Solution Answer Base 1. Describe the potential problems which Inter-European might face in drawing up a sampling frame for the research study. How might Inter-European deal with each of these potential problems? (20 marks) The relevant module is Module 4. Potential problems with the sampling frame lie in any deviation which may exist from one to one correspondence between the customer list used and the population of credit card customers. For example, the list used must be up-to-date (for example, regarding income status) and not omit any members of the population. Ineligibility (i.e. listing individuals who are
31

not/no longer members of the population lapsed customers) is a further potential problem. Other problems may include duplication (holding more than one credit card) and clustering (joint account holders). A good student answer will consider possible methods of dealing with each of these problems. 2. Discuss the benefits and drawbacks of Inter-European using stratified and cluster sampling methods to determine the relationship, if any, between socio-economic characteristics and credit card usage amongst their credit card customers (20 marks) Stratified and cluster sampling are both probability sampling methods so sampling error may be stated in mathematical terms. In stratified sampling, the population is separated into sub-groups called strata. The division is mutually exclusive and exhaustive which means that every population element is assigned to one stratum only and no population elements are omitted in the assignment procedure. Separate simple random samples are drawn from each sub-group. In cluster sampling, the population is separated into sub-groups called clusters and a sample of clusters is drawn. The element which distinguishes the procedures is that with stratified sampling a sample of elements is selected from each sub-group, whereas in cluster sampling a sample of sub-groups is chosen. In the case of Inter-European, it may make sense to separate the population of credit card customers by country, e.g. credit card customers in UK, France, Germany, etc. Using stratified sampling, samples of credit card customers will be drawn from each country. This will allow comparisons to be made between countries. Using cluster sampling, a sample of countries will be chosen as the focus of the research study. A good student answer will note that stratified sampling can improve the costeffectiveness of a research project as this technique allows sampling error goals to be achieved with smaller sample sizes than are required in simple random sampling. A consequential reduction in the total cost of the research can result under certain conditions. Cluster sampling can also be cost-effective. Despite having higher sampling errors than simple random sampling of equal size, cluster samples allow a large enough increase in sample size to more than offset their inefficiency so that overall sampling error is reduced for any given budget.

32

In the case of Inter-European, cluster sampling may be the logical approach if analysis of credit card spending patterns by country is not part of the survey objectives. Cluster sampling will deliver cost savings in terms of interviewers required by country and questionnaire translation expenses.

December 2008 Selected Student Answer

Essay Solution: Inter-European 1. Describe the potential problems which Inter-European might face in drawing up a sampling frame for the research study. How might Inter-European deal with each of these potential problems? The banking network of the Inter-European Bank is extremely large and widely dispersed. Therefore, conducting such a research study will not be an easy task and certain constraints need to be overcome. The Inter-European Bank wants to draw up a sampling frame for their research study. This sampling frame is basically a list of the population, that is, their customer base. However, unfortunately, no sampling frame will ever be perfect, and there will always be an element of error. The difference between the actual population, and the one that is used (the working population/operational) will always be different, no matter how cautious one is. Reasons for this are due to the information used. Lists of a population are essentially outof-date on their publication day as changes are constantly happening. Telephone numbers and addresses will change with move-outs and move-ins and peoples marital, educational and income will change too. This all impacts on their credit card usage, as their lifestyles change to new circumstances. Also, in lists, it is quite common to find duplications and omissions of data. So maybe, using two lists instead of one for the sampling frame will assist Inter-European in narrowing this gap. It is important that measures are taken so that such a gap is reduced, otherwise their research can bring about misleading results. Two elements that will also contribute to increasing error are random sampling error and non-sampling error. Random sampling error occurs when the chosen sample is not a perfect representation of the required population. This will evidently occur due to the less than perfect lists and sampling frames used. Non-sampling error is made up of response error and non-response error. Non-response error occurs when people do not reply; and this may occur quite frequently. People may be very difficult to get a hold of or else will refuse to participate in their research study. Such individuals are also generally different in some way from those who reply. Those who do reply may induce error by providing them with socially acceptable and pleasing answers, or else may forget and provide the wrong information.
33

Inter-European could inform their customers either by a personalised letter, or a telephone call beforehand, to inform them that their participation is to be required in a research study. This can substantially aid participation. 2. Discuss the benefits and drawbacks of Inter-European using stratified and cluster sampling methods to determine the relationship, if any, between socio-economic characteristics and credit card usage amongst their credit card customers A stratified sample is a form of random sample, meaning that each and every element of a population has a known non-zero chance of being selected to participate in a sample. Such results can then be projected to the entire population, and sampling error can also be estimated. A stratified sample is most appropriate when used with large consumer populations, which is precisely the case with Inter-European. The members of the population (its customer base) are divided into mutually exclusive groups to reflect specific social and economic backgrounds. Random samples are then taken from each of these groups. The benefits for Inter-European with this method are that they will have an equal, extensive and detailed coverage of their whole customer base. The customer base will effectively be divided into its specific socio-economic characteristics and credit card usage patterns, so their behaviours will be easy to analyse and this will make it easier for them to target. The drawback is that they may not be able to assess the likely importance and impact of such a group. It could also be that particular members may be categorized according to their extensive and repeated card usage; but the amounts (deposited and withdrawn) may be very small and not compare to those are categorised as less frequent users; but involving larger amounts. A cluster sample is also a random sampling method. Here, specific areas that are known to include a lot of their members are assessed. People in the particular area are targeted as many of their customers live there. This is beneficial as they know where to locate them and target them in future. However, still, little is known about the usage patterns. Are many of their card holders in that area students, or high powered lawyers or executives? And what are the individuals attitudes and perceptions towards the usage of their credit card? Is it used for special occasions large purchases or for frequent every day use? Stratified and cluster sampling methods do provide the opportunity to have their customer base neatly categorized and classified by specific characteristics, but usage patterns, as we have seen, can be quite tricky to determine. Other research must complement such a study in order to get more variable and insightful information.

34