Have an idea?

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

Irregular price intervals

We are doing a brand - price study where I have 20 brands and another attribute of price with levels p1 - p5.

We are showing price using conditional formatting. My concern is that price is not equally spread, levels of brands are -20% -10% 0% 10% 20%, some are -18%, -15%, 0, 13% 15%, some are -25% -10% 0% 15% 20% etc

Do you foresee any problem in utilities computation due to this irregularity? Would you prefer to setup the study such that there are 20 attributes with 5 levels each - alternate specific approach?
asked May 4, 2017 by anonymous

1 Answer

0 votes
Often when we do conditional pricing tables, we need to model interaction effects to explain the data.  If you make the price intervals different in your conditional pricing table, the likelihood of finding significant interactions of course increases.

Modeling as main effects + interactions between SKU and price under conditional pricing tables is mathematically identical in the utility estimation procedure as using alternative-specific price attributes per each SKU.
answered May 4, 2017 by Bryan Orme Platinum Sawtooth Software, Inc. (174,440 points)