The A/B test we refused to run

I know how that sounds, especially here. A/B testing is sacred in our world. You test every screen, every CTA, every copy variant, and you ship what wins. That's the playbook. Most consumer apps run it religiously, and learning apps run it the hardest, because engagement is the metric that pulls in users, investors, and reviews.

We ran one A/B test early on and stopped. The reason is simple. The thing that wins the engagement test, almost every time, is the thing that makes the user do less work. A bigger button instead of a paragraph. A tap instead of a prompt to write. A streak instead of a real reflection. A summary instead of the source. These all "win." Every single one.

And every single one, made the actual learning worse.

There's a Harvard study I keep coming back to. Students in passive classrooms felt they were learning more than students in active ones, while objectively learning less. The feeling of having learned and actually having learned are two different things. The products that scale optimize the first, because that's the metric on the dashboard. So we made a decision that still feels a little uncomfortable. We optimized for outcomes instead of engagement.

In Welica:

• No streaks. No badges. No leaderboards.

• Every lesson ends with a real prompt to think and write about, not a click.

• Concepts come back days and weeks later through spaced refreshers, because that's how memory actually works.

• The exercises are a little harder than they need to be. On purpose.

The risk is real. The engagement A/B test would have shown us things we left on the table. The early funnel is probably slower than it could be. People used to optimized-to-the-millisecond apps will notice the friction. But the friction is the feature. The hardest part of learning is also the part our products have been working hardest to remove. The small, effortful work of trying, getting it wrong, writing it down, finding the gap in your own thinking, and trying again. That work is what changes the person doing it. You cannot tap your way to it.

We launch on Product Hunt tomorrow. If you've ever wrestled with the engagement-versus-outcomes tension in your own product, I'd genuinely like to hear how you thought through it. Where did you draw the line?

And if you want to try Welica, an upvote tomorrow means more than I can say.

Built for becoming, not performing.