ANALYZE THIS
A Guide to Methods of Analyzing Data

By Glenn Iwata

Part III: An Overview of Multiple Regression, Perceptual Maps and Conjoint Analysis

Multiple regression:
Relationships, drivers, and models — Sounds like a bad movie from the 50's. WestGroup often uses multiple regression to model reality. These statistical models can help determine which attributes have the greatest impact on customer satisfaction, advertising awareness, or revenue. A regression model can also help you determine how satisfaction, awareness or revenue could be affected by changing an attribute (e.g., reducing wait times, or increasing the advertising exposure).

Example: Customer satisfaction model
What drives satisfaction? The client assumed that their superior customer service overcame their high prices. Unfortunately, the model identified that this was only true for a small segment of their customer base. Price was eroding customer satisfaction, but was being compensated by their variety of products and unique product options.

Perceptual maps:
If a picture is worth 1,000 words, a perceptual map is worth at least 50 cross-tabulations. Some people would rather see an answer, than read about it. Perceptual maps help illustrate several concepts simultaneously in one picture. The most common request is to map how the client compares to several competitors on specific attributes and then contrast the ratings against the relative importance the customers place on each attribute.

Example: Product position map
How does Zowie Company compare to their top five competitors? Zowie thought their customers most associated them with providing high quality and a creative curriculum and this helped Zowie set their pricing strategy. However, the map helped illustrate that three other competitors were almost equally strong in high quality and creative product positions. The map further helped them to look at other strengths that did differentiate them from their competitors.

Conjoint analysis:
Few decisions we make in life are cut and dry (except harvesting wheat). Any time we make a purchase decision, we are taking into account the variables such as price, color, benefits, value-added services ... and we compare them against our choices —- Mercedes or Lexus, AT&T or MCI, chocolate silk pie or jello (Okay, not all decisions are difficult). A conjoint analysis helps model the choice process consumers go through when simultaneously placing value on various attributes of the product or service options. Would I pay another $1,000 for the warranty package or $500 more for the fancy rims?

Example: Product Price and Packaging
WestGroup recently conducted a study to evaluate packaging and pricing options for a health and beauty product. A conjoint study among current users and potential customers helped to identify the best price point for the product as well as the most important attributes they should highlight on their packaging.

This is the third installment of a series of articles explaining various techniques available for data analysis. We invite you to visit our web site at www.westgroupresearch.com to review discussions and examples of significance testing and market segmentation.

If you are interested in learning more about how to apply the appropriate statistical techniques to your organization, please feel free to contact any WestGroup Analyst at 602-707-0050.

FEBRUARY 2001 Answers:
Home | Previous | Next
Past Issues