How Long Should You Run a Landing Page Test For

Test duration decides whether a landing page result is real. Stop too early and randomness looks like a winner; run too long and outside changes muddy the comparison.

Most guides say run for two weeks. That number is not magic. The right length depends on your traffic, the size of the effect you expect, and the rhythms of your audience.

Why two weeks is a rough default

A full week matters because weekday and weekend behavior differ. A test that runs Monday to Friday may show a lift that vanishes when casual weekend visitors show up on Saturday.

Two weeks covers two of each day type, smoothing out daily swings. But it is only a floor. If your sample is small, you may need far longer to reach a verdict.

Signs a test ran the wrong length

• You called it on day two because one version jumped ahead.
• You let it run for months while the page around it kept changing.
• The sample is tiny but the numbers look decisive, which usually means luck.
• You never set a stopping rule, so the test ends whenever someone remembers it.

A test with no planned end does not measure your page. It measures how long your team stayed interested.

Set the duration before you start

Decide the sample size and the end date in advance, then do not touch the dial until both arrive. This removes the temptation to stop at the first good news, which is the most common way teams fool themselves.

If a genuine emergency forces an early stop, label the result as inconclusive rather than calling a winner. An early peek can inform a hunch, but it should not set your default page.

Key takeaways

• Two weeks is a floor, not a rule.
• Cover a whole number of weeks to balance weekdays and weekends.
• Set the sample and end date before launch.
• Treat early stops as inconclusive, not final.
• Do not extend a test while changing the page.