With every new test, there is a new possibility to learn. If our test hypothesis is proven correct, we can use this info to roll out site-wide changes. If the hypothesis is faulty, we’ll tweak the test to procure better results.