Price testing methods are the ways a business gathers evidence about how customers respond to different prices, from surveys and interviews to live experiments with real purchases. Each method trades speed, cost and realism differently, so the right choice depends on the decision and its risk.
What people say versus what they do
Price tests fall into two families. Stated methods ask people what they would pay or which option they would choose. Revealed methods watch what customers actually buy at different prices.
Stated methods are fast, cheap and safe, because no real price changes. Their weakness is that people often say one thing and do another when money is involved. Revealed methods are more realistic, but they take longer, need enough traffic and carry some commercial risk. Many teams use a stated method to narrow the range, then a live test to confirm.
Survey-based methods
- Van Westendorp price sensitivity meter asks four questions: at what price the product feels too cheap, a bargain, expensive and too expensive. The answers outline an acceptable price range.
- Gabor-Granger asks whether a respondent would buy at a series of prices, which builds a simple demand curve.
- Conjoint analysis asks respondents to choose between product versions that combine features and prices, revealing how much each feature is worth to them.
- Monadic testing shows each respondent only one price, so answers are not shaped by comparison. It has its own entry.
Surveys work best for new products, new markets or large changes, where a live test would be costly or slow.
Read survey answers as a range rather than a single price. They show where customers start to hesitate and where a low price starts to raise doubts about quality, which narrows the options for a live test.
Live tests with real purchases
- A/B testing shows different prices to randomly split groups at the same time. It has its own entry.
- Before-and-after tests change the price for a set period and compare results with the period before.
- Regional or store tests change prices in some locations and keep others as a comparison.
- Offer tests keep the list price but vary the discount, bundle or shipping threshold.
Showing different prices to different customers at the same time can raise fairness concerns and, in some places, legal questions. Rules vary, so check them before running this kind of test, and be ready to honour the lower price for customers who ask.
Choosing a method
- For a product that does not exist yet, use conjoint analysis or Van Westendorp.
- For a small change on a busy product page, use a live test.
- For a shop with low traffic, use a longer before-and-after test or a survey.
- For a business with several stores or regions, use a regional test with matched comparison areas.
- When the risk of upsetting customers is high, prefer offer tests over list-price tests.
Planning a test before it starts
- Write the question the test must answer, such as whether a 10 percent increase keeps profit per visitor level or better.
- Choose one product or product group and one change.
- Decide the success measure and the minimum number of orders before starting.
- Pick dates that avoid holidays, promotions and known competitor events.
- Record everything that happens during the test, including stock-outs and site changes.
- Agree in advance which result would lead to a permanent change.
A written plan stops the team from reading the result in whatever way suits them afterwards. It also makes tests repeatable, so a result found in one category can be checked in another.
A worked example
For example, assume a subscription box sells for 30 dollars, costs 18 dollars to fill and ship, and converts 4.0 percent of 10,000 monthly visitors, which is 400 orders. The team raises the price to 33 dollars for four weeks. Traffic stays at 10,000 visits and conversion falls to 3.7 percent, or 370 orders.
Revenue rises from 12,000 to 12,210 dollars. Gross profit rises more, from 12 times 400, or 4,800 dollars, to 15 times 370, or 5,550 dollars. The result looks positive, but a before-and-after test cannot rule out seasonal effects, so the team repeats the comparison against the same weeks of the previous year before deciding. All figures are assumed.
Reading results honestly
Run a test for at least one full weekly cycle, and longer for products bought less often. Decide the measure of success, usually profit per visitor rather than conversion alone, before the test starts. Small samples produce noisy results, so a difference of a few orders should not drive a permanent change.
Outside events can spoil a test. If a competitor cut its price halfway through, your result reflects that cut as much as your own change. Checking competitor price history for the test period helps confirm that the market stayed stable.
Common mistakes
- Trusting survey answers as if they were real purchases
- Stopping a live test as soon as the result looks good
- Measuring conversion and ignoring margin
- Testing during a holiday or promotion and treating the result as normal
- Testing several prices at once with too little traffic to tell them apart
Key takeaways
- Price testing methods are either stated, based on what people say, or revealed, based on what they buy.
- Surveys such as Van Westendorp, Gabor-Granger and conjoint analysis suit new products and large changes.
- Live tests include A/B, before-and-after, regional and offer tests.
- Choose the method by traffic, risk and how big the decision is.
- Judge results on profit, over a full cycle, and check for outside events such as competitor price changes.
Frequently asked questions
What is the most reliable price testing method?
Live tests with real purchases are usually the most realistic, because they measure behaviour. They still need enough traffic, a sensible duration and protection against outside events.
How long should a price test run?
At least one full weekly cycle, and longer for products bought less often. The test should also reach enough orders that a difference between prices is clearly larger than normal variation.
Can small shops run price tests?
Yes, though with less traffic they may rely on longer before-and-after tests, offer tests or short customer surveys rather than split tests that need many visitors.
Related terms
In BlackAha, Smart Price helps you get a recommended selling price for every product, based on competitor and market prices.