ANALYTIC TOOLS

Travis Research gets the most out of data using a spectrum of analytic techniques. Although some have complex statistical underpinnings, the output presented is readily grasped by clients through straightforward analysis and graphics. A short list of what’s in our toolbox:


ANOVA

ANOVA (Analysis of Variance) goes beyond simple t-tests, which determine the statistical significance of differences between two groups, to test for significant differences between multiple groups simultaneously.
(Return to Top)

CHAID

CHAID (Chi-squared Automatic Interaction Detection) identifies the characteristics that most distinguish one group from another. As an example, one might want to determine what differentiates those respondents who are purchase positive from all others. CHAID splits the sample, first by the one variable that best distinguishes between the two groups, and then goes on to successive splits that are progressively less discriminating. This is readily grasped in graphic output that takes the form of a “tree diagram.”
(Return to Top)

Cluster Analysis

This technique organizes marketing research survey items, or the individuals responding to those items, into groups called clusters, such that the items or individuals falling in any one cluster are similar to one another but different from those falling in the other clusters. Cluster analysis, often in conjunction with factor analysis, is extremely useful for identifying segments of customers having unique attitudes, behaviors and characteristics.
(Return to Top)

Factor Analysis

Factor analysis is often used to find patterns of responses to a marketing research survey. For example, ratings of companies on perhaps dozens of attributes can be reduced through factor analysis into a few key dimensions of perception. Names are then assigned to these dimensions based on their constituent attributes (e.g., price/value, product quality, customer service, technological innovation, etc.).
(Return to Top)

Perceptual Mapping

Perceptual mapping provides insight into how consumers view a brand or product relative to competing alternatives. Several advanced multivariate techniques can be used for this purpose, including multi-dimensional scaling, factor analysis and correspondence analysis. All of these techniques produce a graph or map in which brands or products are positioned on dimensions of attitudes (e.g., quality, price/value, etc.). Those brands/products perceived by consumers to be similar to one another are located close together while those seen as dissimilar are located far apart.
(Return to Top)

Preference Mapping

This technique takes perceptual mapping a step further. Using the same advanced multivariate techniques in combination with regression analysis, preference mapping produces graphs that array brands or companies on dimensions of attitudes that predict purchase interest.
(Return to Top)

Pricing Evaluation and Sensitivity: PRICE-MAX

This proprietary Travis Research technique utilizes responses to four simple survey questions to identify optimum price points for both existing and new products, prices that can maximize unit sales and profitability. PRICE-MAX also generates estimates of demand throughout a broad range of potential prices.
(Return to Top)

Regression Analysis

This multivariate technique is most often used to identify the drivers of key consumer attitudes or behaviors. A typical application of regression analysis would involve identifying just those few out of many perceived attributes of a product that actually predict purchase interest or behavior.
(Return to Top)

Trade-off Analyses

It can be argued that most purchase decisions involve trade-offs. In choosing among alternatives, consumers may take several factors into account. In a new car purchase, these could include exterior design, interior design, horsepower, seating capacity, fuel economy, quality/reliability perceptions, brand name and, of course, price. Two techniques are available to explore and understand these decision-making factors, conjoint analysis and discrete choice modeling. The two techniques differ somewhat in terms of what the survey respondent is asked to do, but their output is essentially similar. That output not only allows the marketer to determine the relative importance of the various factors in the consumer’s decision, but also lets them play “what if” games, essentially “tweaking” various factors, e.g., price, to determine the likely impact on purchase interest and/or market share.
(Return to Top)

TURF Analysis

TURF (Total Unduplicated Reach and Frequency) Analysis is a technique that can determine such things as:

  • The optimum media mix (e.g., the selection of magazines), the mix that maximizes the total number of people who will be reached at least once rather than perhaps reaching a much smaller number of people multiple times.
  • The constellation of product features or benefits that should be created, offered and/or emphasized that will maximize the number of potential buyers rather than motivating potentially far fewer buyers with multiple features or benefits they may find appealing.

A unique Travis Research option, Weighted TURF, can take into account the cost to the company of purchasing each of the media being considered, or providing each of the product features/benefits that could be offered, so that overall expenditures can stay within a pre-determined budget.
(Return to Top)