I recently re-launched my company website and as part of any new site development, knew that I should be doing some obligatory content variant testing. To be honest it seemed a sure fire way to justify/back up our decisions and changes, keep the wolves at bay. We’d done A/B testing on smaller items before and hadn’t seen much difference. This time it was a big surprise, assumptions that we’d made were actually proved wrong – so my attitude to A/B testing is now one of respect and slight awe.
So we had our homepage design but couldn’t decide whether to prioritise customer logos above product information. Some people said ‘visitors want to be instantly wowed by who we work for’ and others weren’t convinced and thought they should be there but below the fold and product information.
A/B tests are very simple to set up using your Google optimizer account. Simply Google ‘website optimizer’ click on the link and log in to your Google account. Choose your type of test (A/B experiment is a good place to start) and it will take you through a step by step set up process.
You need to make sure that you have web pages set up for each variation, for example www.webtest.com/product and www.webtest.com/product2.
You also need to identify a conversion to measure by whether that’s making a purchase, submitting a form, visiting a certain page.
Once you’ve completed the set up they will give you a code that needs to go on your website and once live you’re ready to start testing!
Picking the winner
Throughout the test you will be able to see which of your variations is winning, Google will tell you if there is enough of a difference to select a winner or if more time is needed. If a clear winner is found then it will let you know and you can stop the test – sometimes you will find that there is no clear winner and you just need to accept that it doesn’t actually matter which version you use!
If you still don’t feel ready to start A/B testing your website content then check out this great website (www.whichtestone.com) that does a new test each week. You can vote on a winner and see if you’re correct. It’s a good way to see how other people are testing and also make yourself aware that what we assume is not always the correct answer!