Your Churn Problem Is Probably Hiding in the First Session

You see retention fall off a cliff after a few weeks. Your gut says people are leaving because your product is not good enough. So you spend money winning them back with emails and discounts. It rarely works, because you are treating a symptom.

The cause usually sits earlier, in the gap between signup and the first time someone actually gets value. Fix that gap and the retention curve moves on its own. Here is how to think about it and what to do with your team.

Count the people who actually finished setup, not everyone who signed up

Start with how you measure. If your retention report is built on signups, everyone who never got their account working counts as churned. That makes your churn look catastrophic and points you at the wrong fix.

Claudiu Murariu saw two companies both sitting at around 10% retention, certain they had a churn crisis. When he pulled out only the users who finished onboarding, retention was 50% after a week and 25% after a few weeks. Same product, very different story. The people who got set up were sticking around. The signups who never got going were dragging the whole number down.

Watch activity churn, not just payment churn. People stop using the product before they cancel, sometimes months before. Activity is your early warning, and it is easier to save someone who is drifting than someone who already quit.

Find the one action that predicts they stay

Once you separate the two groups, find the moment where value clicks. For that SEO tool, retention only settled once a user launched three campaigns. Below three, they wandered off. That number became the target: get people to three campaigns and they stayed near 25%.

To find your version, don't guess. Natalie Marcotullio lays out a clean way to pick it in her five ways to improve activation rate: list the key actions in your product, keep the list to ten or fewer, pull six months of usage data, find the actions taken by 40 to 50% of users, keep the ones that happen fast, and pick the one that lines up with conversion. For reference, she cites Lenny Rachitsky's survey showing an average activation rate of 34% and a median of 25%.

One warning. Correlation is not cause. Jane Portman puts it plainly in her activation funnel guide: if all your good users visit the Settings page, that does not mean shoving people to Settings will help. Use judgment, not just the spreadsheet.

Onboarding runs past the first screen

Think of onboarding as the whole stretch between signup and a genuinely engaged user, not the welcome tour. Conor Dewey splits it into two phases in his principles for new user onboarding: the first session, and the follow-up engagement after. Lose someone in the first session and you likely lose them for good.

So work backward. Figure out why people signed up, then map the shortest route to the value they came for. Jane Portman calls this the minimum path to awesome. Every extra step, every unclear button, every integration that looks scary is a place people quit.

Remove obstacles along that path. Preview the payoff before the setup work so people know why they are bothering. And end the first session with a small win, something that proves the product does what they hoped, so momentum carries into day two.

Watch real people get stuck, then fix one line at a time

The fastest way to find friction is to watch someone use your product cold. Michal Mazurek spent hours on screen shares watching new users, and the results were humbling. Things obvious to him, the person who built the app, baffled real users.

He found volunteers just by asking in Facebook groups and Slack channels. His rule during a session: shut up and take notes. If he talked more than 10% of the time, he was doing it wrong. When a user asked how to do something, he asked them how they would try, instead of answering.

Most fixes were tiny. A reworded sentence. An example next to a box. Not big features, just hundreds of small edits. After 15 rounds of this, his activation rate jumped from 42% to 74%. The lesson he landed on: nobody reads your instructions, so show, don't tell.

The deep cut

Not every retention problem is an onboarding problem, and Murariu is careful to say so. Sometimes people finish onboarding, hit the value, and still leave a few weeks later. That points somewhere else. In one case, small businesses churned because they only ever needed one campaign, not three. The product worked fine. The company was selling to the wrong customers.

This matters because it stops you from over-fixing onboarding when the real issue is who you let in the door. Before you rebuild your flow, check whether the people churning were ever a fit. A smoother onboarding cannot save a customer who does not need what you sell. Separate the two, or you will polish the funnel and still watch the wrong people leave.

Three questions for your team

• When we report retention, are we counting from signup or from the people who actually finished setup? Rerun the number both ways this week and see how far apart they are. That gap tells you if you have an onboarding problem.
• What is the one action that, once someone does it, makes them stick? Pull six months of data, keep the shortlist under ten, and name your version of the three-campaigns number so the whole team aims at the same target.
• When did we last watch a brand-new user set up the product without helping them? Book three screen shares, stay quiet, and turn what they trip over into a list of one-line fixes.