Conversion 101: Setting Up Your Test

Once you've created and validated your hypothesis for optimizing your conversion rate, you can finally move on to the testing phase of CRO! I know we've spent a lot of time talking about double-checking your assumptions before getting to implementation, but I think that a "measure twice, cut once" philosophy is very valuable here.

As a quick recap, here's what we've covered so far:

• Using Best Practices to Generate Hypotheses

• Validating Hypotheses

• Deciding Whether Testing is the Right Method (vs. surveys, interviews, etc.)

So now it's time to pick one hypothesis and start setting up your test! Remember, it's important to only test one thing at a time - you may be excited about all of the conversion improvements you could make, but the only way to get a clear result is to isolate a single variable.

Once you've decided what to test, you can either set up an A/B Test or Multivariate Test:

• A/B Test (Split Test): A/B Testing is best when you are comparing two different layouts of the same page. You will create two versions of the page and show each one to 50% of visitors. Whichever version has the best conversion is the one you should make permanent!

• Multivariate Test: Multivariate testing is best when you have multiple sections of a page that you want to change. You will create a page version for each variation of these changes, and then show each one to an equal % of visitors. Whichever combination of changes has the best conversion is the one you should make permanent!

Once you've created your variants, you will decide who to target (we recommend targeting all visitors for basic tests) and how long to run the test (between 2 & 4 weeks depending on site traffic). If you're using Google Optimize, here are the instructions on how to set up a test in the tool. Now you've officially got a conversion test!

Bennett