3 ms·
I think you misunderstood, or maybe I wasn't clear in my post. This wasn't an experiment on 2 separate groups; this was an experiment on the same group: those t
by ezekg 2y ago
I think you misunderstood, or maybe I wasn't clear in my post. This wasn't an experiment on 2 separate groups; this was an experiment on the same group: those that convert from trial to paid. When I did X, conversion rate for the entire group increased, i.e. I made more money. If I stopped X, I'd make less money. That then lead me to the point of the post -- the % difference in conversion rate were the people who weren't actually ready but signed up for a trial anyways, only to run out of time and never come back.
- ang_cire 2y agoOk, I think I see what you're getting at: You're assuming that the percentage increase in conversions after you extended or removed the trial timeout period across the board corresponds to users who previously would have been kicked out by the trial expiry before they could see the value in converting. My comment was referring to the period you mentioned where you were offering people manual extensions, and you noted that you saw more conversions among that group than among users who didn't extend their trials.
- ezekg 2y agoThat's right. I acknowledged that that increase in that comparison was likely biased since those people would also add a card, which increases conversion rate in itself (but also weeds out those that are uncommitted). But that doesn't change the fact that overall conversion rates were higher when I went out and manually offered extended trials.