How to test a price change before you launch it

Most price changes are not tested. They are announced, then watched nervously. The teams that do test often test badly: several changes at once, on everyone, for a week. This guide is the sequence we use with teams at Precheck. It works whether you simulate first or go straight to the test.

1. Frame the decision as options

Write down the options you are actually weighing and what you would do differently depending on the answer. "Should we raise the price?" is not a decision. These are:

• Bundle at ₹1,499, bundle at ₹1,299, or keep selling singles at ₹549.
• Starter plan capped at 3 seats or 5.
• Annual-only default for mid-market, or keep monthly.

If you cannot name the option you would pick under each outcome, you are not ready to test. You are ready to think.

2. Explore by segment before you test

A price rarely lands the same way on every customer. In our illustrative simulations the pattern repeats: repeat buyers respond to bundles because they are already stocking up, first-time visitors will not commit to three units before trying one, and lapsed customers barely respond to price at all because price was not why they left.

Before spending on a test, explore how each segment might respond across the price points. Methods that work at this stage:

• Van Westendorp price sensitivity surveys, still the most used pre-test instrument, as SaasDash's guide notes, and cheap to run.
• Cohort analysis of past price moves, if you have made any.
• Data-informed simulation, which is what Precheck does: simulate the response distribution per segment and rank the barriers. We say plainly that our predictive accuracy is not yet established, so treat the output as a way to choose the test, not to skip it.

The output you want is not "the right price". It is "the segment and price where the outcome is most uncertain and most consequential". That is where the test goes.

3. Run one test

The rules are boring and they are the whole game.

• One change. Price, or packaging, or default term. Not all three. Monetizely's guidance is blunt about it: testing multiple elements at once makes it impossible to know which change drove which outcome.
• One segment. The one your exploration flagged.
• A control group. The same segment, the old price, the same window.
• Enough volume. If the segment cannot produce a readable difference in the window, pick a bigger segment or a longer window, not a looser threshold.

4. Measure the thing that hurts

Conversion in week one is the easy number and the least important one. Track:

• Conversion and revenue per customer in the test window.
• Second purchase or first renewal, because that is where a price rise shows up as churn. Rework's 2026 pricing guide makes the same point: short-term revenue lifts are tempting and can be paid back in churn.
• Support tickets and cancellation reasons, read by a person, not a dashboard.

5. Decide, then write down what you learned

The test settles the decision. It also produces the most valuable asset you have for the next one: a decision, an audience, and an outcome. Keep it. Teams that have a few of these can calibrate any simulation tool, including ours, against their own history. That is exactly what we are asking teams to send us.