Ashley Stirrup

There may be a goldmine of growth hidden in your last experiment. Most teams read a test as a single number "+0.9, decent, ship it" and move on. But that one average can quietly bury the most valuable thing the test found. So I'm kicking off a new series: Data Science, Decoded. Our data scientist Hampus Poppius writes brilliant, rigorous breakdowns of how experimentation actually works. They're also… a lot, if you don't live in statistics. So I'm taking his most powerful ideas and translating them into something any PM or marketer can act on today, with no jargon, no formulas. First up: why your average lift is hiding money. Crack that "+0.9" open and it isn't one result, it's a huge spread. Most customers barely moved. A few moved a lot. The average sits in the middle and describes almost none of them. Here's the part that matters: the customer segment that responded the most to the offer are also the ones least likely to accept the free trial. Only a portion of them accepted the offer, so they barely move the headline number, and the average buries them completely. That's the goldmine. Not in the people already saying yes — in the high-upside group that's tuning you out. The move isn't to celebrate +0.9. It's to go dig it out: rework the free trial for them, make it more visible, give them longer to try it, lead with "no commitment." So before you ship on a lift number, ask two questions:  → Who reacted the most to this offer?  → Did that segment frequently accept the offer? When your biggest winners have your lowest adoption, you haven't found a result. You've found a to-do list. The full breakdown is in the carousel. Hampus's complete guide (stats included) is linked in the comments.