Which test won?

I’ve written a few blogs about A/B Split Testing: how it works, how it surprises, why you should do it, etc.

There is a great website – Which Test Won – which documents different A/B tests and invites its users to guess which version of a test won.

From time to time, you get it wrong and you learn something you didn’t know. Or discover that what you thought you knew has been discredited.

Some of the tests are low-level and others are very high-level; the high-level tests can sometimes be frustrating if only that you’d like more drilling down to understand what differences had driven the increase or decrease in performance: when many variables are altered between concurrent pages in the test, it can be difficult or impossible to know which changes were influential.

Sometimes, the answer is inexplicable and simply goes to show that you need to constantly test across your website in order to extract the best possible website performance.