Have an idea?

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

One conjoint or segment specific conjoint

Issue: We have pre-defined segment variable (assume student, professional and affluent etc.) We want to design a new credit card product that appeals to each segment. Although attributes might be similar but levels might be different  - for example, air port longue access, travel reward miles etc. wont appeal so much to student while cash back may not appeal highly to affluents. We are stuck with following dilema:

1. Option 1: Design segment specific multiple conjoint exercise and show according to segment. For example, design a seperate conjoint for students

2. Option 2: Do overall alternate-specific design ignoring segments, run latent class segments and then profile LC segments by our segments.

Any input is highly appreciated.
asked Jun 12, 2014 by Mike

1 Answer

0 votes
Dear Mike,

I don't see much of a dilemma here because I don't your option 2 as a viable solution.  If the levels of attributes varied in accord with some other conjoint attribute then an alternative-specific design might work for you but if they vary by different respondent groups I think you need to go with Option 1:  you need to show your different respondent groups their appropriate attributes and levels.

Now, if you want to verify rather than assume that some attributes are more appropriate for one group of respondents than for others, then you may want to give everyone the same conjoint experiment (though I still don't see the need in what you've written to make it an alternative-specific experiment).
answered Jun 12, 2014 by Keith Chrzan Platinum Sawtooth Software, Inc. (90,475 points)