AB Testing | Process Street AB Testing – Process Street

Introduction

A/B testing which can sometimes be referred to as 'split testing' is great for comparing two versions of web pages to see which one performs better. This checklist is to be run by your marketer/growth hacker who will be comparing two different web pages when you're looking to find out what your conversion rate is. This template can only be used for one test at a time. 

Over at Process Street, we created this AB Testing Checklist to help increase your conversion rate.

Experiment Prioritisation:

Note down experiment details

First things first you're going to have to note down the experiment details of your AB test that you're carrying out. You're going to record the following:

  • The Web pages that you're testing
  • The variable of the element you're testing (headline, images, offers, etc)
  • The Element that you're testing (original headline, new headline, etc)

Do this by using the form fields below.

Expected test length

Next up, you'll want to make sure that you let your test run long enough in order to obtain a substantial result. Otherwise, it'll be hard to tell whether there was a statistically significant difference between the two variations of your elements. If you're unsure on how long the test should be run for, check out this AB test duration calculator.

Record your expected test length using the form field below.

Choose your target

Even though you will be measuring a number of metrics on one test, you should pick your target for this test. You need to know what you're looking to achieve from this test but keep the target realistic. For example; It could be that you want your conversion rate to be 20% for your chosen variable.

Record your original conversion rate and new conversion rate, using the form fields below.

Configuring Test:

Setup AB tool

For the configuring test phase, you will firstly need to set up your AB tool before you begin your test. Also, make sure that your tool works properly when doing so. Do this by using the form fields below. 

Make sure the test works

Next up, you have to make sure that the conversion trigger 'element' is working and the test itself is working. Do this by using the form fields below.  

Deploy test live

Once everything has been tested and setup you're now good to go to deploy your AB test live. When you run A/B tests, you'll need to run the two variations at the same time, otherwise, you may be left guessing your results

Test on live

After the test has been deployed. You now have to test it once again when it has been deployed, to make sure everything is working correctly as planned. It will give you an insight into what other people see and are reading. 

Results:

Record results

Once the test is over and has completed, you now have to record your results from the test. Do this by using the form fields below. 

Take action

The test has been complete and you have your results so now you have to get rid of the losing variation in your AB test. The variation that was statistically better is the winner, so you choose that one!

If your test proved inconclusive meaning that the test was the same for both web pages, run another test to just make sure but if that is exactly the same then stick with your original variation as changing it clearly won't make a difference

Using the form field below, decide if you want to implement the change

Sources:

Relevant Checklists: