Have an idea?

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

CBC/HB - Conditional Pricing estimation of utilities

For one of my attributes, I made use of placeholders (medium, high) and the conditional pricing function to definte certain combinations of attributes as price values ($0, $3, $6). However, when running the utility estimation in CBC/HB, it return utility values for the placeholders instead of the conditional pricing values. Just wondering, how can I get utilities for the conditional pricing values instead of the placeholders, thank you!
asked Sep 5, 2012 by anonymous

1 Answer

+1 vote
With conditional pricing you are still estimating a single utility for each of the placeholders, so if you want to simulate Brand A at the high price point, you would add the utility of Brand A plus the utility of "high" together to get your total product utility.  The respondent would have seem an actual number, i.e. $6, but a single utility is estimated for "high" for that respondent.

If you estimate interaction effects to allow for different price sensitivities, you would end up with a utility for Brand A, a utility for "high" and then an additional utility for Brand A x high that you would add to the total product utility for that combination.
answered Sep 5, 2012 by Brian McEwan Gold Sawtooth Software, Inc. (38,990 points)
if interaction needs to be added to get the utilties for the conditional prices, but the interaction utilties now is only 2x2, how can the interaction utilites be added when the conditional pricing is based upon more than 1 attribute (excluded price)?or more than 2 attributes and so on?
Not sure I understand.  An interaction effect can only be estimated in our software between two attributes.  If you involved more attributes in the relationship, you could recode your design down into larger attributes and estimate, or you could add multiple interaction effects between the attributes involved to try to model things better.  When multiple attributes are involved, however, we strongly recommend that your conditional pricing table be proportional to give you the chance to model things with main effects only.