Have an idea?

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

The higher the ZCD None Utility, the fewer the number of products with positive utilities?

Dear forum,

I would like to create segments of customers according to the number of products with positive utilities (purchase preferred over none option).
Is it correct to use the ZCD none utility as a proxy, when comparing customers?
Thus is the assumption correct that a customer with a high ZCD none utility has automatically fewer product combinations with a positive utility compared to a customer with a low / negative ZCD none utility?

Thank you very much for your reply.


asked Sep 27, 2018 by Chris Berlin Bronze (570 points)

2 Answers

+1 vote
I'm assuming you are using CBC.  In that case, the scaling of the None utility depends on how often the respondent uses the None.  Sometimes, using the None truly represents that the respondent is less likely to choose (buy) something in the real world.  But, sometimes using the None is more related to the positivity/negativity tendencies of the respondent when interacting with the survey.  Sometimes the None usage is more related to the fatigue affecting the respondent.  So, I get a bet nervous when trying to use the None to segment respondents into more or less likely to purchase groupings.

Zero-Centered Diffs scaling is just a normalized scaling of the raw utilities, where a multiplier is found to scale up the magnitude of the parameters so that the average difference between best and worst levels within attributes is 100, per person.  So, the interpretation of the None vs. the sum of utilities for other concepts should be the same no matter whether using zero-centered diffs or the raw utilities.
answered Sep 27, 2018 by Bryan Orme Platinum Sawtooth Software, Inc. (186,865 points)
+1 vote

It is correct that respondents with higher positive values for None will have fewer concepts that exceed that None utility, all else being equal.  

Like Bryan, I'm a little reluctant to attach a hard meaning to None, as if it was a perfect calibration for willingness to purchase.  That said, if you assume that the None measurement isn't missing the boat and that it's really capturing an unwillingness to buy any of the concepts on offer, then I think your interpretation is correct.   

So I'm being a little less conservative than Bryan is here, but I would proceed with some caution.
answered Sep 27, 2018 by Keith Chrzan Platinum Sawtooth Software, Inc. (103,225 points)