Estimating Willingness to Pay
Posted by Vovici Blog on Wed, Jan 26, 2011
David Lyon of Aurora Market Modeling is a regular presenter at the American Marketing Association’s annual Applied Research Methods conference, teaching a class on “Survey-based Approaches to Pricing Research”. He begins his class with a discussion of the most basic technique, but one that survey authors often get wrong: Willingness to Pay (WTP).
First, the questionnaire describes the new product or service in detail, then asks the user, “How much would you be willing to pay for this?”
It’s better not to provide a list of possible prices for the respondent to choose from: make it an open-ended question. A list of prices biases the answers. Dave shared the example of a space tourism survey, conducted before multimillionaires decided to spend tens of millions of dollars to go up on Russian spacecraft. The prices given were $1,000, $2,000, $3,000 and $4,000 – low for what many seem willing to pay (Virgin Galactic will be offering suborbital flights at $200,000).
When analyzing the results, don’t use averages but plot the cumulative percent willing to pay at any given price: for example, if respondents answered $5, $10, $15 and $20, any respondent that answered $15 is also willing to pay $5 and $10.
Willingness to pay is a lousy direct question to ask respondents, who tend to lowball their answers, in effect bargaining rather than answering accurately. Nor, given the artificialness of the exercise, are their answers likely to reflect their actual behavior. As a result, use this technique for those times when you really have no idea what people are willing to pay and treat it as an input into further research. Since it is an input, you can—and should—use a small sample size. Then test the results with other techniques, exploring price points that are above the Willingness-to-Pay answers.
Regarding those other techniques, David presented extensively on the merits of using direct questioning and trade-off methods for pricing research. The Applied Research Methods conference is atypical of U.S. MR conferences – it’s actually a series of 2-hour and 4-hour classes on a range of research topics. This year I’ll be teaching classes on Panel & Community Management and on Pragmatic Social Media Research. You can learn more at the ARM 2011 web site.