Pricing Tests: How to Measure Willingness to Pay

Price is the assumption founders test least and guess most. Yet it decides whether a business works at all. A product that people love at $5 and abandon at $20 is a different company from one they happily pay $50 for, and you can learn a surprising amount about which you have before you write much code.

Why asking "what would you pay?" fails

Ask people directly and you get anchored, strategic or simply invented answers. Some lowball to get a deal; others name a number to be polite. Survey methods built for pricing get around this by asking several structured questions and looking at the pattern rather than any single answer.

Van Westendorp's price sensitivity meter

Proposed by the Dutch economist Peter van Westendorp in 1976, the method asks four questions about a clearly described product:

1. At what price would it be so cheap that you would doubt its quality?
2. At what price would it be a bargain?
3. At what price would it start to feel expensive, but still worth considering?
4. At what price would it be too expensive to consider?

Plotting the cumulative answers gives a range of acceptable prices. It is cheap and quick, but it still measures stated attitudes, so treat the range as a hypothesis.

Gabor–Granger and conjoint

The Gabor–Granger technique, developed by the economists André Gabor and Clive Granger in the 1960s, shows respondents a series of prices and asks whether they would buy at each one, producing a rough demand curve. Conjoint analysis goes further by asking people to choose between bundles of features and prices, revealing how much each feature is worth to them. Conjoint needs more respondents and careful design, but it is the standard tool for products with many configurable options.

Testing with behaviour

• Price on the smoke test. Put a price on your landing page and run two versions with different prices to separate audiences.
• Price in the pre-sale. A real price on a pre-order or LOI is the strongest evidence you can collect.
• Tier experiments. Show different plan line-ups to new visitors and watch which plan they choose, not only whether they convert.
• Discount depth. Offer early-access discounts of different sizes and see where uptake stops rising.

Anchors and decoys

People judge prices relative to what is around them. In Predictably Irrational (2008), the behavioural economist Dan Ariely described a magazine subscription offer with three options: web only, print only and print plus web, where print only cost the same as print plus web. Almost nobody chose the print-only option, but its presence made the bundle look like a bargain, and when it was removed in his classroom experiments, preferences shifted towards the cheaper web-only plan. Your pricing page is always an experiment in anchoring, whether you mean it to be or not.

A price change hidden in plain sight: 6:5 blackjack

The gambling industry offers one of the clearest examples of a price change most customers barely noticed. In traditional blackjack, a natural (an ace and a ten-value card as the first two cards) pays 3 to 2: $15 on a $10 bet. Over recent decades many casinos, especially on lower-limit tables, switched some tables to pay 6 to 5: $12 on the same bet. The game looks the same and the change is printed on the felt, yet it raises the house edge by roughly 1.4 percentage points, which in a game where the edge with good basic strategy can be well under one per cent is a very large price rise.

Casinos effectively ran a long, public pricing experiment: change the payout, watch table occupancy and revenue. Experienced players learned to look for 3:2 tables, but many recreational players kept playing. Table minimums are priced dynamically too, rising on busy weekend nights and falling in quiet hours. The lesson for founders: the price customers notice is not always the price they pay, and the details of your terms (payout ratios, renewal rules, usage limits) are part of your pricing. We go deeper into how casinos design and test their products in casino product design as a case study in testing.

Common mistakes

• Pricing from cost rather than value.
• Testing price with an audience that is not the buyer.
• Treating early adopters' price tolerance as the market's.
• Changing several things at once, so you cannot tell what moved conversion.