The Effect of Advertisements on Consumer Perception

Memorandum

Subject: The Effect of Advertisements on Consumer Perception

Thank you for choosing BP&S Consultants to evaluate the results of your newest ad campaign for the top-selling ball point pens. After reviewing and analyzing the data supplied by Mr. Wong, we recommend that the proposed campaign not be used. We realize that your organization has already invested a lot of time and money, but we would ask that you genuinely consider the consequences of implementing an ineffective campaign: Sales could drop if the wrong ads are used, more money would be spent to create a new campaign after inevitable failure, and your organization's brilliant reputation is on the line and could be tarnished. Our recommendation is based upon statistical tests and is backed by the results.

Mr. Wright encouraged a simplified process of the data analysis believing that the data was collected from a completely random sample. In compliance with Mr. Wright's wishes, we constructed a One-Factor ANOVA Completely Randomized Model (Attachment A) and evaluated. The test keeps things "simple" and also provides accurate results. The conclusion for this test showed that the advertisements did not affect the consumer ratings of the pens.

Mr. Wong expressed concern that the rating scale that was used did not satisfy the criteria of being a true interval or ratio level of measurement [1 (poor) to 7 (excellent)]. Taking this into consideration, we completed a Kruskal-Wallis Test (Attachment B) which allows for the data to be ordinal and therefore tested more accurately for the utilized rating scale. The results of this test also showed that the advertisement campaign did not affect the consumer ratings.

Upon further discussion, Mr. Wong communicated that the data supplied were ratings from 9 people who had read each of the ads and evaluated pens. This newly supplied information, while recalled only vaguely by Mr. Wong, led BP&S in a different direction for testing but did not devaluate any of our previous tests. We took the new data and put it into a Randomized Block Model (Attachment C) which controls for the variation due to personal preference of the surveyed consumers. This test illustrated that the advertisements did affect the consumer ratings.

To identify differences between the Advertisements, a Pairwise Comparison of the Advertisements - Tukey Method (Attachment D) was conducted. This comparison allowed for the conclusion that the only significant difference between ads existed with Ad D. Looking back at the individual advertisements, we realized that Ad D tended to "greatly oversell the pen's characteristics" and performed most poorly out of the campaign.

Lastly, we evaluated the data that was supplied from the second focus group. Each consumer was asked to read each of the five ads and state which they considered to be most convincing. To assess this data, a Goodness of Fit Test for a Uniform Population (Attachment E) was performed. This test showed that the five different advertisements were not equally appealing to this group of consumers.

All tests that were performed lead BP&S to conclude that the proposed Advertisement Campaign is ineffective and should not be implemented. We suggest that the statistical data be presented to QBA Advertising and that a new campaign be created at no cost to your organization. As professionals, we would expect that QBA Advertising would acknowledge the failed campaign and work with your organization to see a new, successful campaign through to fruition. We would also encourage your organization to work closely with QBA Advertising on the next phase of the campaign and to inject creativity and uniqueness into the presentation of your product. Ball point pens are a "dime a dozen" type of commodity... what separates your product from the pack. We thank you for your business and we are optimistic that you will consider our recommendations.

Attachment A: One-Factor ANOVA Completely Randomized Model..............................3

Comparing more than Two Populations - One-Factor ANOVA Completely Randomized Model: Consumer ratings of pens do not differ.

Attachment B: Kruskal-Wallis Test 4

Kruskal-Wallis Test: Consumer ratings of pens do not differ.

Attachment C: Randomized Block Model 5

Randomized Block Model: Consumer ratings of pens do differ.

Attachment D: Pairwise Comparison of Means - Tukey Method 6

Pairwise Comparison of Means - Tukey Method: Advertisement D