Have an idea?

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

How can I deduct WTP from summed price ACBC conjoint analysis?

I have used a summed price ACBC conjoint test to estimate utilities. Now I would like deduct the WTP from these utilities.

When using linear price, I kown that you can convert utilities to WTP by dividing the utilities by the linear price utility (one unique utility from price).

However, now with the summed price I get at least two utilities for price so I am not able to divide all utilities by one single coefficient (price utility) to get a monetary value.

Can somebody assist me on this issue?

Thanks!
asked Aug 30 by anonymous

1 Answer

0 votes
If you run a linear model, you get one price coefficient that represents the change in utility per change in currency unit. With ACBC, we report the endpoints since these are easier to use for most simulations, importance calculations, etc.

If you need the slope of that line, simply divide the change in utility by the change in price to get the utility per currency unit.

For instance, if ACBC returned $10=2.39 and $30=-2.39, then we can compute the utility change per dollar as (2.39-(-2.39))/(10-30) = 4.78/-20 = -0.239u/$. In other words, each $1 increase in price reduces the utility by 0.239 utility points.
answered Aug 30 by Aaron Hill Silver Sawtooth Software, Inc. (7,535 points)
If you want the $ value of each utility point, just flip the equation and divide the change in price by the change in utility. You'll end up with the $ change per utility point.

For instance, if ACBC returned $10=2.39 and $30=-2.39, then we can compute the dollar change per utility point as (10-30)/(2.39-(-2.39)) = -20/4.78 = $4.18/utility. In other words, each utility point is worth $4.18.

For more information on interpreting WTP, see https://www.sawtoothsoftware.com/support/technical-papers/general-conjoint-analysis/assessing-the-monetary-value-of-attribute-levels-with-conjoint-analysis-warnings-and-suggestions-2001.
Yes, it is. See my post from just a moment ago where I reference that paper.
...