A strong answer. “Our onboarding completed at 34%, and the assumption in the company was that it was too long — seven steps — so the brief I was given was 'make it shorter'.

I pushed back on that brief for a week, which is the part of this project I'd defend hardest. I watched twelve session recordings, and people weren't dropping out from fatigue — they were dropping out at step three, which asked them to connect a data source. They didn't have the credentials. They weren't the person who had them.

So the problem wasn't length. It was that we'd designed a flow for one person when the job needed two.

What I designed: step three became skippable, with a 'send this to your admin' path that generated a link. Everything after it worked in a sample-data mode so the product was usable before the connection existed. That's a harder build than deleting steps, and I had to argue for it.

Completion went to 61% over six weeks. But the number I care more about is that week-two retention went up more than completion did, because the people who now got through were people who could actually use it, not people we'd pushed through a shorter funnel.

What I got wrong: my first version of the admin hand-off email was written by me, and it was terrible — a 4% click rate. Our lifecycle marketer rewrote it and it went to 19%. I should have brought her in at the start.”