About 90% of the founders we talk to think they have a retention problem. So they ship more features, spend more on ads, hire another SDR, tweak the dashboard nobody uses. And the bucket keeps leaking. The part almost nobody looks at is the one part of the product every single user goes through: the space between signing up and becoming a paying customer. That space has a name. It's called activation — the moment a new user actually experiences the core value of your product, the "okay, I get it, this is for me."

Here's the uncomfortable version: no amount of sales or marketing can save a bad product experience. Marketing brings traffic, eyeballs, exposure. That's its job. But if the product doesn't serve users when they arrive, churn is more or less guaranteed. A customer is someone who pays — and to get from stranger to user to paying customer, the product has to deliver value first.

The three reasons new users don't convert

In our work we see the same three patterns. One, they don't understand what the product does for them — the connection between what your marketing promised and what they see in the app just isn't there. Two, they understand it but never feel the value. Reading that a tool "saves you hours a week" is different from feeling it. Three, the onboarding flow itself gets in the way — usually a long list of questions someone on the team thought might be useful, asked before the user has gotten anything back. Company size, department, "invite your colleagues," all before the first win.

How to find where you're actually leaking

Start with the cheapest thing: go through your own product as a brand-new user. Open an incognito window, sign up with a fresh email, and write down every step that slows you down or asks for something you shouldn't need yet. Then pull your product analytics (Mixpanel, PostHog, Amplitude) and find the exact steps where people drop off — the same funnel view we use to debug growth with the AARRR framework. Layer in session recordings (Hotjar, Microsoft Clarity) to see where they hesitate or rage-click. The pattern you want is a step where the drop-off is much higher than the steps around it. That's a leak. Then read your support tickets and negative reviews — users tell you where they got stuck if you read carefully.

Analytics tell you they left; they don't tell you what they were thinking. For that, talk to two groups: users who signed up and never came back, and users who stuck around. Reach out close to the action while it's fresh. Getting happy users on a call is easy; for the ones who left, an incentive like an Amazon gift card works well. Then find what your best users did in their first session that churned users didn't — the point where the two paths diverge is your activation pattern, and the step right before it is what you redesign first.

What to actually change

Cut everything from signup and onboarding that isn't earning its place. For every field, ask: can I collect this later, after the user has felt the value? Usually the answer is yes. Then map the shortest path to the moment the product clicks — finishing a first edit and exporting it, getting a first meeting transcribed and summarised — and strip out everything in between. Guide users with a tour, checklist, or empty-state prompts that point to the three or four actions that lead to that moment, not every button in the nav. And don't stop at signup: the first seven days in the inbox are where users decide whether you matter.

This isn't a one-time fix. Your product and users change, so treat onboarding like any other part of the product — something you keep testing and improving. Sealed from the inside, the leaky bucket produces a different set of numbers: higher retention, lower churn, less support load, lower CAC. Take one product we worked with — after reworking onboarding, signup conversion went up roughly 20%. Same product, different first ten minutes.

If users are signing up and never coming back, that's exactly the work we do. Get in touch and we'll look at your specific flow and tell you what to fix first.