TRC

By Rajan Sambandam

Background and objectives

In order to establish the value of our proprietary, artificial intelligence-based key driver technique, Satiscan™, for consumer products brands, TRC conducted a study among members of TRC’s in-house panel. A study was fielded, in consultation with major consumer products brands, that evaluated brand image attributes for a toothpaste and a laundry detergent brand. The impact of the brand image attributes upon overall purchase interest was then assessed for each brand.

Notes on methodology

The research was conducted among panelists who had purchased and used toothpaste or laundry detergent within the past three months. A total of 681 respondents completed the toothpaste evaluation, while 600 respondents completed the laundry detergent study. Each respondent was asked to indicate their level of interest in purchasing the toothpaste brand and/or the laundry detergent, and to evaluate each brand on a series of image-related statements.

• Overall purchase interest was measured using a five-point scale: definitely would, probably would, might or might not, probably would not or definitely would not purchase.
• Reactions to image-related statements were captured using a ten-point anchored scale in which 10 equals: describes the brand extremely well and 1 equals: does not describe the brand at all.

A ten-point scale was used for the attribute ratings for two reasons:

1. A ten-point scale is easily recognized and used by most consumers, and
2. It was expected that the additional data-points available within a ten-point scale (when compared to, say, a five-point scale) would add to the overall sensitivity of the readings.

Two models of purchase interest were produced for each brand: one using traditional multiple regression; one using Satiscan™, TRC’s proprietary means of key driver analysis. The bulk of this case study focuses on a comparison of these two models. But first, we present for the reader a brief description of Satiscan™.

A brief description of Satiscan™

Satiscan™ is an analytic method developed by TRC specifically to address questions such as, "What attributes are key drivers of purchase interest?" Using artificial intelligence, Satiscan™ examines each respondent’s ratings, learns the patterns inherent in these ratings, and produces a map of purchase intent. This map can be used to identify those aspects of brand image that have a direct impact on purchase intent, those that have an indirect impact on purchase intent, and those that have no noteworthy impact on purchase intent.

This is a significant development in key driver analysis. Traditional methods (typically regression analysis) can only identify direct impacts on purchase intent, and ignore the fact that aspects of brand image inevitably interrelate with each other, and hence have indirect as well as direct effects on consumers’ overall purchase interest levels.

Using artificial intelligence, Satiscan™ reviews all possible path models in relation to consumer responses and only then determines the ideal model of purchase intent. This may not seem like a tremendous advantage until one realizes there are literally thousands of possible path models to choose from, and the odds of choosing the optimal one in advance are extremely slim, unless there is precise theory to guide model building.

Comparing the Satiscan™ models to traditional regression models

The difference between Satiscan™ and traditional regression analysis is clearly shown by comparing the maps developed for the toothpaste and laundry detergent brands. On pages five and seven can be found the models of purchase interest developed via traditional regression. As the arrows indicate, for both brands, these models suggest that three specific image attributes significantly influence purchase interest directly, and that none of these three relate to one another. The numbers (or weights) above each arrow denote the relative importance of each attribute. To simplify this discussion, it is safe to assume that the larger this number, the greater that attribute’s impact on purchase intent.

To read the rest of this article and have access to the entire TRC white paper library, please click here.