How long should I run my landing page A/B test?

Don't rely on calendar time

Running a test for two weeks tells you nothing meaningful. What matters is sample size—how many total conversions you've collected, not how many days passed. A test with 200 conversions across variations is more reliable than one with 20 conversions running for a month.

Calculate the conversions you need

As a baseline, aim for at least 50 conversions per variation before drawing conclusions. If your page converts 2% of visitors, that means roughly 2,500 visitors per variation. With lower conversion rates, you need more visitors.

Account for weekly patterns

If your business is slower on weekends or has weekly patterns, run the test through at least two full weeks to smooth out daily variation. One week of Tuesday-Sunday data doesn't capture Monday performance.

Watch for early stopping bias

The biggest mistake: stopping the test when one variation "wins" after a few days. Random variation can create false leaders. If you see one version leading after three days, keep running. The lead often reverses.

Use statistical significance, not gut feeling

Your A/B testing platform should show confidence levels. Don't stop until one variation shows at least 95% confidence of being better than the other.