3 ms·
Is it possible that the reason they’re still doing it is because it works, for better or worse?
by dahart 5y ago
Is it possible that the reason they’re still doing it is because it works, for better or worse?
- bopbeepboop 5y agoI think it’s much more likely they just don’t measure the bad: Their tests are short duration, so they measure the smaller but more immediate bump — then fail to notice long term user disengagement. It’s common for A/B testing to “switchback” downhill like that: recording mild positive signs in the short-term, highly focused trials while causing long-term diffuse damage to your brand/engagement. Companies are datablind like that.