How long should your conversion optmisation tests last?
Guide to CRO Test Duration Calculator
There’s no such thing as a stupid question!
If you’re trying to decide how to approach your CRO testing programme you may have dozens if not hundreds of tests you would like to perform. You may have already looked at how to order these by priority and you’re all ready to get started, improve performance and add value to the business you’re working with. However what many do not think of is the potential time a test will need to reach statistical significance, a relevant confidence level or in the case of Google Optimise a suitable probability score. This calculator provides would be CRO testers with an idea of how many days or user sessions are needed to achieve a potential uplift in conversions.
Simply go to Google Analytics or another reliable tracking source and take the current conversion rate of your website. Input the desired improvement in conversion rate you would wish to see (we know not all tests are around just conversion rate and we’ll get on to this later!). Then take the number of sessions you get per day. Enter a statistical confidence you wish to achieve and then how many variants are in the individual test you are running and press calculate. Easy (and maybe a little obvious)!
The higher the difference between conversion rates the more obvious the statistical significance and evidence the null hypothesis is false (that the variant won’t make a difference if run again and again). Therefore the less sessions and days are needed to reach that point of statistical proof. The smaller the incremental increase the more data is needed to reach a confidence level or high probability the variant/variants will beat the original.
When adding sessions to the CRO test duration calculator remember it is NOT total site sessions. Instead it is the number of average daily sessions that the individual page/path/test original, receives.