Pricing Sensitivity Model (Van Westendorp)

The Van Westendorp Price Sensitivity Meter is fifty years old and still the fastest honest answer to "what should this cost?" — but most readouts are someone squinting at where four lines seem to cross. This skill computes the crossings: cumulative curves built from the actual responses, intersections found by linear interpolation, non-monotone respondents dropped and counted.

Required Inputs

• Survey responses — per respondent, the four classic answers as prices: too cheap (quality suspect), cheap (a bargain), expensive (getting dear), too expensive (out of the question). 20+ valid responses for a stable read; the script warns below that and refuses below 5.
• Segment splits (optional) — the tool doesn't segment; run it per segment and compare, which is usually where the real finding is.

If the survey hasn't run yet, produce the four questions verbatim and the screener instead, then stop — don't invent responses.

Output Format

1. The four points — OPP (optimal price point: too-cheap × too-expensive crossing), IPP (indifference: cheap × expensive), and the acceptable range PMC–PME. Each with one sentence of meaning, not just the acronym.
2. Data hygiene — valid n, dropped non-monotone count (a high drop rate is itself a finding: respondents didn't understand the category or the questions).
3. The recommendation — a price inside the range with reasoning; note that OPP minimises purchase resistance, which is not the same as maximising revenue — premium positions price above OPP deliberately.
4. The caveat — VW measures perception, not demand; pair with a real willingness-to-pay test before betting the pricing page on it.

Programmatic Helper

This skill ships scripts/van_westendorp.py — zero dependencies (stdlib zip+XML):