Have an idea?

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

Share of preference is still higher than the base case at highest price level, what is best?

For my WTP simulation I have 3 price levels, namely 335, 345 and 355 Euros.
While simulating the base case, the share of preference for the product was 17.1%. After adding the 4 most desired features considering my HB, the share of preference was 29.1% for my highest price level; 355 Euros. Because 355 Euros is my highest price level, how can I reach this base case share of preference of 17.1% to determine the exact WTP? If not possible, what should be concluded from this simulation?
asked Jan 24, 2019 by anonymous

1 Answer

0 votes
I'm assuming you are using CBC.  So, first steps would be to look over your data to make sure you haven't made a mistake of some sort.  Look first at counts, making sure that the levels for attributes seem to have counts that make sense.  Better levels should carry higher count percentages.

Next, you should look to do any data cleaning.  If some respondents were not paying attention or answering lazily/randomly, then this can cause problems for price sensitivity measurement vis-a-vis other attributes.  Depending on the source of the sample, some researchers need to delete as many as 30 or even 40% due to unmotivated respondents.  I typically delete around 5 to 20%.

Next, I worry deeply about WTP calculations.  They are problematic and often overestimate WTP.  That's why you may be "running out of room" on the high point of your price level scale.

Unless respondents are properly recruited and properly motivated to give realistic answers as if they are spending real money, then WTP can be too high (the problem you may be seeing).  Furthermore, simulations like you are setting up may not reflect the reasonable competition and substitutes that are available in the marketplace.  You haven't told me about the static (held constant) products in your simulation.  Do they reflect a reasonable range of available alternatives in the marketplace that represent the majority of the competition?  Have you included the None concept as well as the "outside good"?

Again, I discourage the idea of trying to estimate WTP.  There are many pitfalls to be overcome.  We have written an article outlining some of the problems at: https://www.sawtoothsoftware.com/download/techpap/monetary.pdf

We also have a chapter that speaks more about the mechanics and problems of estimating WTP in our book, "Becoming an Expert in Conjoint Analysis" that is available for purchase (if you live in the USA) from Amazon.  If you are outside the USA, you can order a PDF version of the book from our website at: https://www.sawtoothsoftware.com/products/pricing-ordering#cabook

answered Jan 24, 2019 by Bryan Orme Platinum Sawtooth Software, Inc. (176,815 points)