Background
The Coca-Cola Company recently conducted some studies showing that their most loyal consumers are getting older, and that the younger segment of the population is not drinking enough cola to sustain the growth the company expects for the next ten years. Coca-Cola is also aware of Pepsi’s efforts to attract the young population by hiring popular celebrities like Shakira and Britney Spears for their advertising. Given this information, Coca-Cola has decided to launch a marketing campaign aimed at young people, ages 15-22, who are beginning to make conscious purchases and have potential to become attached to one brand of cola in the upcoming years. Coca-Cola knows how much opinion influences young people’s decisions to buy a product. Thus, as a marketing strategy, they plan to use a blind test to show that most people objectively prefer Coca-Cola over their main competitor, Pepsi. They have hired us to conduct this blind test and provide them with detailed results.
Methodology
The experiment was conducted in a class of 46 students in Georgetown University. The Coca-Cola Company wishes to know the preference of the subjects based strictly on the taste of the colas. The following controls were put into place so that the experiment’s validity was as high as possible.
Double Blind Test – To prevent subjects from stating their preferences on any reason other than taste, the tester served the two colas in identical paper cups. This was done to control for possible existing preferences towards a logo or brand name. Furthermore, the colas were served into the cups behind a screen so that not even the individual distributing the cups to each participant knew the exact content of each cup.
Contamination control – To avoid the carryover effect caused by one cola being ingested prior to the other one, every participant was provided with a salty Ritz cracker which they partly consumed before drinking from each cup.
Counterbalancing - Counterbalancing was cleverly used to prevent further carryover effects and sequence effects. For instance, the numbering of the cups was set up so that half of the subjects tasted Coca-Cola first, while the other half tasted Pepsi first. Also, the survey was formatted in four slightly different variations so that both the colas’ brand names and the questions had different orders. Both of these controls prevented the opinions of those seating next to each other from influencing their neighbor’s answers.
Cup labeling - The paper cups were simply labeled with two consecutive numbers. This simple labeling eliminated potential biases that could have arisen from more leading labels like letters or colored cups.
Neutral Questionnaire – All the premises in the questionnaire were very simple and unbiased. Additionally, when measuring the degree of liking of the product, a simple 9-point liking scale was used for both types of cola.
Results
Once the blind test was finished the subjects’ answers were entered into a computer where statistical software allowed us to evaluate the data in more detail. First, a chi-square test was performed which showed that 33 out of 46 participants labeled the colas correctly and that this same amount reported that they preferred Coke over Pepsi. The statistical results gave us a chi-square figure of 8.696 with a p-value of only .003, which means that the results did not happen by chance and, thus, are significant. Next, a within-subject t-test for subjects’ established preferences for colas was conducted which revealed that Coke, with a mean of 6.5217 on the 9-point scale, was preferred over Pepsi which had a mean of 4.8478 on the same scale. With t(45) = 4.036 and the p-value for the test being less than .001, we can conclude that the means differ significantly from one another. This test was followed by a t-test on the actual taste test. Although the results in this analysis were not profoundly different, in this test we saw the phenomenon of regression towards the mean, since the mean liking for Coke and Pepsi on the 9-point liking scale was 5.7826 and 4.9130, respectively. With t(45) = 2.265 and a p-value of .028, we can again conclude that these results are statistically significant. Finally, a within-subjects t-test was conducted for those participants who expressed a strong established preference for one cola over the other but failed to label the two colas correctly. This test had means of 6.1250 for liking least and 4.7500 for liking most. With t(7) = 3.667 and a p-value of .008, we can conclude that the test is significant and that these participants’ pre-existing preferences for a particular cola were not really based on taste, but instead on previously acquired perceptions of image, quality or brand name.
Marketing Implications
The results from this experiment show that Coca-Cola is the preferred brand of choice of most consumers, based on both taste and pre-existing preferences. However, Coca-Cola should conduct several more blind tastes tests to corroborate these results before launching an advertising campaign that focuses on the outcome of the taste test, since the mean liking for Coca-Cola was just .86957 higher than the mean liking for Pepsi. Thus, another test where more people are involved may show a different mean liking for both Coke and Pepsi. Nevertheless, it is very clear from the test results that Coke is the pre-established brand of choice of most customers, a result that should make the marketing team of Coca-Cola extremely happy. This pre-established preference means that they have created a strong brand name and a superior image of quality which is carried by most consumers in their minds. In our opinion, Coca-Cola should continue to exploit their brand name and superior image when targeting young consumers, and perhaps forget about the actual taste difference between colas. Coca-Cola has not had an aggressive ad campaign in a while, and possibly that is all they need to attract the young consumer segment to their already popular product.