knowing it's working

Praise is cheap; coming back is not. The difference between a product people like and a product people need is retention, and almost every other number can be faked or bought.

Retention is the one unbiased signal I’ve found. You can inflate signups with incentives, buy a launch spike, pull a hundred thousand wallets from a points program — and none of it means anything. A hundred people who return next week without being paid means everything. Someone pitched me two hundred and fifty thousand users and I wasn’t interested; someone else had five hundred users and a hundred who came back every week, and I was very interested. Two thousand users at thirty percent weekly retention beats a hundred thousand at three percent, every time.

So find the core loop — the one action people repeat — and watch whether they repeat it. That action is the product; everything else is decoration. The honest test is a reason to come back without a notification. If a happy user has no reason to return next week unless you poke them, the loop isn’t built yet, and no amount of good session time saves you. A well-funded app once showed me beautiful thirty-three-minute sessions and a retention floor of zero. Great numbers, empty bucket.

Measure it from day one. Install PostHog or Google Analytics before you launch, put UTM tags on everything, and look at one number every week — retention. You’d be shocked how many funded teams run campaigns with no attribution and call it intuition. If you stare at the same number every Friday, your subconscious starts optimising for it, so make sure it’s the right one.

And don’t put projections next to it. What’s working is what already happened — the cohort that came back, the curve that bent. Certainty about a number you haven’t earned yet is a red flag, to an investor and to you.