Stopping a test too early gives false winners. This rule-of-thumb estimate helps new and soon-to-launch businesses plan how long a test should run.
Results
Visitors needed per version
Total visitors for two versions
Approximate weeks to run
How to read the result
The estimate uses a common approximation for 80 percent power and 5 percent significance. Low-traffic pages need bigger changes to reach a clear result.
Practical tips
Decide the sample size before starting
Run tests for whole weeks
Test bold changes on low-traffic pages
How our team helps
Knowing the theory is one thing; having specialists apply it every week is another. Our strategists, SEO and advertising specialists, designers, developers and AI engineers do this work for new and soon-to-launch businesses across Canada, with one plan and one point of contact, led by Ali Sedighi, MBA.