Unit 5 · Maximizing Performance Through Iteration · Lesson 5.2
How to test a price
A price test needs long enough for guests to see the rate, consider it and book. Two weeks is the working cycle. Success is two or three bookings at the price you advertised, not one, because a single booking can be an outlier. The most common error is ending a test early and calling the result evidence.
The full lesson text below is an edited transcript of the video, published 2026-08-24. The complete course is free at the playbook.
How long a test needs
Start from your booking pace, because it sets what is even observable. A listing chasing full occupancy on two and three-night stays needs roughly three bookings a week to hold it. Most listings do not run at full occupancy and their stay lengths vary, some one night, some seven, so the arithmetic moves around. What stays true is that a couple of days produces too few bookings to mean anything.
Two weeks is the recommended testing cycle. It gives the market time to see the new rates and respond, favourably or not at all, which a two, four or five-day window does not. It is also short enough that you are not burning unsold calendar while you wait.
What a successful test looks like
A test succeeds when guests book your listing at the rate you advertised. One booking is not that. A single booking can be an outlier: the right guest at the right moment, telling you nothing reliable about the price.
Two or three bookings at a given price point is the threshold worth trusting. At that point you have evidence the market accepts the rate, rather than evidence that one person did.
When to stop early
The two-week cycle has one legitimate shortcut: a flood of bookings. Ten bookings in the first three days is strong evidence the price is too cheap, and there is no reason to keep selling nights at that rate for another eleven days to confirm it. End the test and raise prices.
Absent that, let it run the full two weeks. A quiet first week is not a failed test; it is an incomplete one.
The mistake to avoid
Ending tests prematurely is the most common failure, and it is seductive because an unsold calendar feels like an emergency. Two operators can be equally close to the answer while only one keeps going; the one who stops early never finds out they were nearly there. Unless the bookings are flooding in, hold the cycle.
What you do at the end depends on what the fortnight showed. Sometimes the evidence demands a real change, as with the ten-bookings-in-five-days case where prices are clearly too low. Sometimes it says leave everything alone. As a strategy gets closer to optimal the adjustments get smaller and less frequent, which is itself a signal that the work is converging rather than a reason to intervene harder.