Add target condition or percentage "lift" to A/B test success criteria

Right now just achieving a statistically significant increase in a conversion rate doesn't really account for the desired amount of "lift" we might want.

A small increase in a value may not actually be enough to want to make changes on the back of.

This target lift should give us enough information to calculate a target sample size before the test starts if we're able to query the current baseline via elastic. This sample size should allow us to estimate a recommended amount of time to run the test for.