Most A/B Tests Don't Need More Traffic
The growth team paused four tests after three weeks with no significance. Traffic wasn't the problem — their read on what counts as a real difference was.
Angle
Experimentation
2 posts
Signups rose 30%. Revenue fell 12%.
The A/B test found a lower price that converted better. It converted the company into a worse business, and the tool reported a win the whole way down.