Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Calculating market shares from preference shares

Dear Sawtooth,

After analyzing results from cbc conjoint study for insurance company (car insurance). We had 2 attributes -  price and brand. Option None was also included.
Questionnaire also contained quesiton about currently used car insurance company.
Client is interested in market share estimates based on preference share.
Could you please advise based on your experience how to proceed with these estimates. Probably we can compare current market share and preference share and make a conversion index. However is this too simplistic and could it cover different price levels?

thank you
asked Nov 4, 2019 by Teodor

1 Answer

0 votes

There are a lot of factors the separate respondent's preference shares in a survey experiment from real market shares out in the world.  For example, respondents might be more familiar with some brands than with others, or some brands may have agents that are more accessible to a respondent than do other brands, or some brands might bundle home and auto insurance.  

Our Choice Simulator allows a user to incorporate some of these factors, for example if you know about between-brand differences in awareness and distribution.  

But a straight correction factor will usually not suffice to match preference and market shares, and even if you use the Choice Simulator to account for differences in awareness and distribution, and even if you did so at the respondent levels, you still wouldn't be accounting for the possible effects of bundling or volume discounts.
answered Nov 4, 2019 by Keith Chrzan Platinum Sawtooth Software, Inc. (102,700 points)