After running an experiment for a few weeks and then implemented changes, I realized my sample was not big enough. Now I want to look over my data before/after the implementation to see what % conversion I had and how it changed.
Since the retained users would effect my conversion rates after implementation (had I only viewed it based on time periods), I created cohorts for both time lines before/after my experiment then plugged it into a funnel:
cohort a) Those who signed up before experiment
cohort b) Those who signed up after experiment
I then created a funnel leading to conversion starting at a time period since 30 days before experiment till 30 days after, then I added both cohorts together into the same breakdown.
Does this sound like the right implementation?