# How to calculate the WTP

Here some utilities from my logit model as an example:

Distance
0-50 km    = 0,57950
50-150 km    = 0,03412
150+ km    = -0,61290

Price
500 €    = 0,47323
1.000 € =  -0,04796
1.250 € = -0,42527

Price linear  = -0,45554

I already calculated as an example (because i've read it somewhere): Distance 0-50km/ price linear = -((0,57950)/ (-0,45554)) = 1,27

So it means that I have a WTP = 1,27€?
If you can give me another method I would be very grateful. Thank you!!

Sophie,

A few things here.  Assuming your utilities are correct, the fraction would be the reciprocal of yours (price in the numerator) so that you have euro per km.

Then you want to use the the number of km in one utility point as your denominator, NOT, as you have, the utility of the most appealing distance.

Ideally you would do this at the respondent level, not the aggregate, so you'll want to run a hierarchical Bayesian logit utility model (or a mixed logit) so that you can run this formula for each respondent.  And then you probably want to look at the median (not the mean) of this respondent-level measure.

Finally, for this to work well, we need to know the amount of utility peer km.  Unfortunately, you seem to have measured your distance as ranges (level 1 and 2) and as an open-ended range (level 3).  So is the utility of 0,5795 for 0 km or 30 or 50?  We don 't know.  And what is the number of km  measured by 150+?  Is it 150 km or 200 or is it 384,400 km, the distance to the moon?  Again, we don't know.  Given this inability to tie specific distances to your distance levels, I think your data is not well-suited to computing a WTP.

Alternatively, you could look at the amount of euros it would take to equalize the difference between 0-50 km and 150km+, which is 0,5794 - (-0,6129) or 1.1923.  That euro difference is the equalization price, another measure of WTP.  Again, doing this at the respondent level and using the median is probably your best bet.  This gives you the amount of euros the median respondent would pay to compensate them for the difference in your maximum and minimum distances (but it will not directly give you a euros per km metric).
answered Jan 14 by Platinum (77,750 points)
Keith has provided great advice here.  In our introductory training on conjoint analysis (like we'll be offering on Jan 28-30 in Florida), we stress that conjoint levels should be written so that they have concrete meaning for respondents.  To use squishy language or to provide ranges rather than specific amounts as you've done leaves room up for interpretation by each respondent.  It hinders your ability to interpret what the data mean.

I'd like to point out an article we wrote on the subject of WTP: https://www.sawtoothsoftware.com/download/techpap/monetary.pdf

And, Keith and I extended this article to include more information on WTP in a chapter in our book, "Becoming an Expert in Conjoint Analysis".

With typical marketing applications of WTP, a key thing to remember if you pursue the algebraic approach to estimating WTP is that you are ignoring the likely presence of competition.  If a person intrinsically is willing to pay \$100 for something, this may not mean that a firm could actually charge \$100 to this respondent if a competitor comes along and offers a suitable substitute at a price of \$10.
Dear Keith,

thank you for your fast answer. Is it possible that you can have a look at my data? I've made an HB estimation but I have still trouble in understanding how to calculate the WTP. I would like to send you my excel sheet of the HB estimation and maybe you can calculate an example for me.  since english is not my mother tongue and I'm not the best in math, things are getting difficult for me. Thank you so much. Please contact me on my email address if you like.
Sophie, you can send your file to me at keith@sawtoothsoftware.com