Pricing Research Methods: A Practical Comparison
7 methods to find your price. Only 2 actually predict market behavior. Here's how to choose.
Why pricing research fails before the market sees a price
Conjoint analysis says customers will pay €X. The market launches at €0.7X — and even then, churn spikes.
Or: Van Westendorp identifies a price range. Marketing settles on the upper end. Volume collapses.
Most pricing research fails for the same reason most consumer research fails: it measures stated willingness-to-pay, not revealed willingness-to-pay. The say-do gap hits pricing harder than any other research domain — because the gap between what consumers say they'll pay and what they actually pay is almost always biased downward.
The methods below differ in how seriously they take that gap.
Seven methods, ranked by predictive validity
1. Van Westendorp Price Sensitivity Meter
4 stated-preference questions about price perception ("too cheap", "cheap", "expensive", "too expensive"). Outputs a price range.
2. Gabor-Granger
Ask respondents at increasing price points whether they'd buy. Builds a price elasticity curve.
3. Conjoint Analysis
Respondents trade off product features (including price) across choice scenarios. Output: utility scores per feature.
4. MaxDiff / Best-Worst Scaling
Forced trade-offs between feature/price combinations.
5. Behavioral / Implicit Testing
Reaction-time-based measurement of price-value associations. Measures gut-level acceptance, not deliberate calculation.
6. A/B Pricing Tests in Market
Actual purchase data at different price points across matched consumer segments.
7. Causal AI for Price-Decision Drivers
Causal modeling on existing market and behavioral data to identify true price-elasticity drivers.
"Stop using Van Westendorp alone. The methodology costs more in lost margin than it costs in fieldwork."
Which method should you use?
| Your Situation | Best Method |
|---|---|
| Quick directional exploration, low budget | Van Westendorp (with caveats) |
| Single-product price optimization | Gabor-Granger + behavioral validation |
| Multi-feature product launch | Conjoint + Causal AI validation |
| Premium positioning where stated WTP underestimates | Implicit testing |
| You can A/B test in market | A/B testing — period |
| Strategic price increase or decrease | Causal AI |
| Cost of being wrong is high | Implicit + Causal combined |
What Frank recommends — and what he doesn't
If you only have budget for one method: it's almost never Van Westendorp alone.
If you only need a directional answer: Van Westendorp plus a smell-test from someone with category experience.
If the pricing decision matters strategically: Implicit + Causal. Anything less is gambling with confidence intervals.
Frequently asked questions
What is the best pricing research method?
There's no universal "best" — it depends on the decision. For high-stakes strategic pricing, the combination of Implicit measurement + Causal AI typically beats stated-preference methods (Conjoint, Van Westendorp). For quick directional work, Van Westendorp can be acceptable with awareness of its biases.
How much does pricing research cost?
Stated-preference methods (Van Westendorp, Conjoint): €15–50k. Implicit testing: €30–80k. Causal AI for pricing: €50–150k depending on data complexity. A/B testing in market: variable (usually cheapest if technically feasible).
When should you use Van Westendorp vs. Gabor-Granger?
Van Westendorp: when you want a price range and four reference points. Gabor-Granger: when you want a price-elasticity curve. Both have stated-preference limitations.
Does conjoint analysis predict actual purchase behavior?
Mixed evidence. Conjoint correlates with behavior in well-designed studies of low-involvement product categories. For premium positioning, sensitive products, and strategic pricing decisions, conjoint underperforms vs. implicit + causal methods.
What is willingness to pay (WTP)?
The maximum price a consumer would pay for a product or service. Methodologically: stated WTP (from surveys) is different from revealed WTP (from market behavior). The gap is often 20–40%.
How does causal AI change pricing research?
Causal AI identifies the causal drivers of price acceptance — not just statistical correlations. This matters when you're not just predicting current willingness-to-pay, but deciding how a price change would shift the demand curve.
Get a pricing-research diagnostic
Most clients come to us because their existing pricing research said one thing and the market did another. We help you avoid that, before it happens.
Get my AI Diagnostic →