I was asked the question recently, if there’s an ideal number of visitors for a conversion rate test, and if is what is that, and here is my answer:

The thing is, to be statistically significant (or to get statistically significant results) in many cases your significance level will be dependent on the number of results, not the amount of traffic (or sample size). (Sample size does matter, but for non-scientific tests like variable traffic conversion rate tests, results are more important than sample size, in my opinion).

I like to work with the conversions, not the # of visitors. Because the number of conversions is what helps you get statistically significant results. So…I like to see 30-50 conversions per option per campaign. So if you are testing a squeeze page with average conversion rate of 30%, then you might only need 100-150 visits per campaign, yielding 30-45 conversions. But if you are selling something that has a conversion rate of for example 1%, then you will need 3000-5000 visits.

For initial testing purposes, I recommend going with lower prices or squeeze page prices (free) to get results for the various parts of your sales page. THEN test the price with the highest converting sales page.

The reason for this is that if you are testing much higher prices, then your conversion rates will of course be much lower, which means that the test will take a lot more time. So speed up the test by reducing or removing the price barrier, then once you have significant results, put the price variable back in (or raise it to reasonable levels for your test).