Another extraneous factor is maturation, which involves external changes unrelated to the experiment, which affects subjects’ response (McDaniel and Gates, 2002). In this case financial situation of the customer at the time of experimentation could be a factor. For example, if they have just received their salary they may have a more disposable income.
As a field experiment, and taking into consideration extraneous variables the experiment seems to have a higher degree of external validity.
Was the convenience store a good venue for the experiment?
Having discussed external validity in the previous section this provides a basis to judge the advantages and disadvantages of using the convenience store in this experiment. As it is conducted in an actual environment the design has certain feasibilities. Since the convenience store is comparatively small it easy to observe and control the whole procedure. Therefore the costs should be relatively low and results easy to collect.
On the other hand some shortcomings need to be considered. The chosen suburban area may not realistically take into account the characteristics of differing locations in the real world, for example the consumer structure may differ between suburban and metropolitan areas. The differences between convenience stores and supermarkets must also be considered. People do convenience purchases in convenience stores and regular purchases in supermarkets, therefore it can be concluded that peoples purchase behaviour varies between the two. Supermarkets will provide a wider variety of brands of beer to customers. That is to say, Dixie may face must stronger competition in supermarkets.
How effective was the marketing design?
Having discussed the advantages and disadvantages of the experiments, the graph in appendix 1.0 illustrates and compares the sales of the seven brands of beer following the introduction of the M.A.P.S plan.
Overall Dixie indicates the highest sales along with Bud and Miller respectively, the lowest sales being Dixie Light and Coors Light. The hypothesis of the experiment was that the location of brands on a display would affect sales. However, the graph clearly shows that there must be external factors affecting the results because in general it caused the sales to decrease or remain constant. On average Miller, Bud and Dixie Light decreased by 1.7%, while Coors and Coors Light remained constant. There were two exceptions to this trend, Dixie and Miller Lite, where sales increased. However, Dixie sales increased by five times more than that of Miller Lite, resulting in Dixie being the overall best seller.
Before taking into consideration the effect of the location of the beers, external influencing factors other than shelf positioning must be considered. Such factors could include: previous advertising; price advantage; packaging design and taste, all of which would have an impact on the consumer and therefore affect the hierarchy of sales of Dixie down to Dixie Light.
Considering the effect of the M.A.P.S. treatment no definite conclusions can be drawn from the results. In the case of Dixie and Miller Lite, change in location of the brands did increase sales. This however defeats the object of the exercise as the change in position of Dixie Light has decreased sales by 1% whereas Miller Lite, the main competitor, noticed an increase in sales. This shows that the M.A.P.S. treatment did not produce its desired outcome. Dixie sales dramatically increased, whereas sales of Miller decreased showing brand position may have an effect. However, in a study of this nature you cannot rely on one factor alone to determine consumer decisions. A more reliable and valid outcome could be achieved by including tests for other factors in the research design, for example timeframe, weather, taste, advertising and promotions.
How could the marketing Design be improved?
Taking these factors combined with the disadvantages into consideration improvements could be made to the M.A.P.S. plan to provide more interpretable results. Firstly the problems of the original experiment have to be addressed. The most important of these is timescale. Treatment one was conducted in the first two weeks of April and the second in the last two weeks. This had an effect on external factors, such as weather affecting purchasing, promotions in other stores, and increased advertising of other brands, and therefore the experiments were conducted in different conditions, showing that concomitant variation does not occur as it cannot be proven that shelf position is the only factor affecting sales. To increase the accuracy of the results treatment one and two should be conducted at the same time in controlled environments. Therefore both treatments will be subject to any changes in external factors that may occur making the results more comparable. The experiment would therefore have to be held in two stores of similar size with the same set up in terms of catchment area, size, shelf space and turnover, for example two supermarkets within a chain could be used. Also, by using a supermarket the number of sales and customers will be increased as people are more likely to make such purchases when conducting their weekly shop.
By placing the brands in the same environments, this indicates that any changes to sales will be due to the factors under experiment rather than any others. Two groups will be required; a control group which consists of the original style of shelf positioning and an experimental group consisting of the M.A.P.S style. This experimental design can be represented symbolically as shown below.
Experimental Group: O1 X O2 (O = observation)
Control Group: O3 O4 (X = treatment)
The effect of the treatment can be estimated by (O2 – O1) – (O4 – O3). This will show the effect repositioning has on beer sales, with external factors constant for both groups making it a fairer test than the M.A.P.S plan. The effects of the extraneous factors can be illustrated by (O4 – O3 ), (McDaniel and Gates, 2002).
As mentioned before Dixie Light experienced the lowest sales among all brands whether its shelf position changed or not. It is therefore necessary to find more detailed information about their products. By conducting a general questionnaire, non – specific to the ‘Dixie’ brand for security reasons, it could enable the researcher and company to identify other factors that may need to be tested, for example, price, awareness of the product. Ideally the questionnaire could be conducted on the shop floor in the beer aisle to ensure the correct audience is targeted. It will have to remain short to keep customer attention and be in multiple choice format making it quick and easy to answer. In addition to this, if the Dixie Brewing Company have their own website, they could incorporate a more personalised questionnaire to discover customer views on their products, particularly with reference to taste of the products. Alternative products could be produced and taste tests carried out, however this would be an expensive process so pretests and prior research will have to be carried out to sum up cost benefit analysis. From the results of the questionnaires, it is possible to determine what influences consumer decisions, therefore some changes can be made to not only the research design but also the product.
The alternative research design outlined above will provide more reliable results, therefore making the experiment more internally valid, because as far as possible, other influencing factors have been removed.
References:
Churchill, G.A., Iacobucci, D., (2002), Marketing Research, Methodological Foundations, Eighth Edition, New York, Harcourt College Publishers.
Malhotra, N.K., (1999), Marketing Research An Applied Orientation, Third Edition, London, Prentice-Hall International (UK) Ltd.
Mc Daniel, C., Gates, R., (2002), Marketing Research, The Impact of the Internet, Fifth Edition, New Jersey, John Wiley & Sons Inc.