Have an idea?

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

DCE with one category/ attribute


we would like to conduct a dce on incentives. Question is: Of the following alternatives which would motivate you the most to join our program?
Alternative A: incentive_1a ($25 in cash), incentive_2a (toaster), incentive_3a (gift coupon for 2 cinema tickets)
Alternative B: incentive_1b ($35 in cash), incentive_2b (microwave), incentive_3b ( gift coupon for 1 cinema ticket)
Alternative C: I prefer not to join the program.

So there are 5 attributes (incentive_1, incentive_2 .... incentive_5) with 3 levels each.
But the point is alternatives are not really realistic. Future customer only get one incentive for enrolling the program.
And as every attribute actually belongs to one category (e.g. all levels like cash, kitchen tools, gift coupons could be part of only one attribute called incentives) probably better method in this case is best-worse case 2. However, we would like to use classic dce for comparison reasons.
Is it possible to perfom a dce with in fact only one attribute (incentives) spread over all alternatives and would we get reliable results?


asked Dec 30, 2013 by Andrew Bronze (1,045 points)
retagged Dec 30, 2013 by Walter Williams

1 Answer

+1 vote
Best answer
MaxDiff (another form of discrete choice analysis) would seem to me to be a better approach to tackle this problem.  These two pieces of research are quite related to your issues and I believe lend support to my recommendation:


answered Jan 1, 2014 by Bryan Orme Platinum Sawtooth Software, Inc. (156,655 points)
selected Jan 2, 2014 by Rajesh Rana
Bryan, thanks for the comment. MaxDiff really seems to be the better method here. And I like the idea of hybrid approach and consider using it for upcoming project.

Concerning above mentioned dce, when next to the existing incentive attributes there will be another attributes relating to costs for customers (distance to next service center, annual costs, minimum time of membership). So one alternative consists of different incentive attributes (benefits for customer) and cost attributes (costs for customer). This makes it little tricky using maxdiff in my opinion as costs & benefits have to be presented together.

Any idea on this. I am kind of stuck here and not sure how to bring benefits & costs together.