Have an idea?

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

Guidance on Van Westendorp PSM and NMS extension results

Hi Team,

We have conducted a Van Westendorp PSM for a new product, along with two additional questions (NMS extension) to derive an approximate demand curve. (The product doesn't have competition, hence the choice of NMS to derive the approximate curve). We have followed these steps to derive the demand curve using Sawtooth's NMS simulator - (https://sawtoothsoftware.com/resources/software-downloads/tools/van-westendorp-price-sensitivity-meter) and manual.

1.  Default probability at different points of purchase intent scale
2.  Half probability at the two extreme price points (too cheap = bargain /2, similarly, too expensive= expensive /2), as we assume there is some uptake of product at these prices
3.  Replicated the calculation for more price points (we have around 60 price points within our testing range to derive the purchase intent and hence we have added more price points using same calculation in the simulator)

As we plug in the data (4 price points and purchase intent data at bargain and expensive price points), we observe that the demand curve is peaking before the acceptable price range, i.e.  the revenue or purchase intent maximizes just before the acceptable price range starts and then it gradually decreases till the last price point. Seems odd. Per our understanding should it not be within the acceptable range

Need your guidance on the following -
- How do we interpret/explain this? Why we are getting this trend?
- Is this trend seen previously, importantly what steps we can take to investigate the data? (weights or different probabilities or else)
- Assuming this trend doesn't change, what should be our recommendation regarding the price of the product? Shall we go with acceptable price range or purchase intent data? (though these seem to conflict each other)

Any thoughts are most welcome.
asked Mar 18, 2021 by Prakhar
Prakhar, I should add that Ming Shan is presenting a paper at our conference next month wherein he shows some MAJOR improvements to the practice of PSM.  Anyone who does a lot of this should probably check out his presentation.  The conference website is https://events.sawtoothsoftware.com/presentations?_ga=2.261955080.449840182.1616253455-1709247883.1563413668

1 Answer

0 votes

To answer your three questions in turn:
- This isn't particularly odd - there is no good empirical reason to believe that the 4 intersection points on the original PSM mean much - which is the reason folks often supplement PSM with the NMS improvement.  When they're in conflict I report NSM and ignore the PSM intersections.

- I have seen this pattern (and many others) previously.  Peak acceptability and peak revenue are inherently different things and all you're doing here is seeing that play out.

- I trust the NMS more than I do the unevidenced PSM intersections.
answered Mar 18, 2021 by Keith Chrzan Platinum Sawtooth Software, Inc. (111,275 points)