Ask a customer who has been with you for years why they stayed, and they will almost never say the setup was smooth. Ask them directly about onboarding and you’ll often get something like: honestly, that part was kind of a pain. Then they’ll tell you, at length and with feeling, about the thing your product did for them three weeks later.
I’ve never once heard a renewal credited to a good onboarding flow. I’ve heard hundreds credited to an outcome.
That’s an awkward finding if you run a trial, because nearly every activation dashboard I open is measuring the first thing and not the second.
Onboarding is friction, no matter how good it is
This is the part people resist, so let me put it plainly: onboarding is not a feature. It is a cost you impose on someone before they get anything. A well-designed onboarding is a smaller cost. It is not a benefit.
Users don’t convert because the cost was low. They convert because something on the other side of it was worth paying. Which means “reduce onboarding friction” — the default activation project at most companies — is optimizing the denominator and ignoring the numerator entirely.
The tell is in the language. Most teams say a trial user is “activated” when they’ve finished setting up: account created, data connected, integration authorized, team invited. Every one of those is a description of work the user did for you. None of them is a description of value the user received.
So the finish line sits in the wrong place, and everything downstream inherits the error. The emails fire when setup completes. The sales handoff triggers when setup completes. The dashboard turns green when setup completes. An entire system congratulating itself for getting someone seated at the table without ever asking whether they enjoyed the meal.
Move the finish line to the first real win
The replacement rule is short: onboarding isn’t over when the user is live. It’s over when they’ve gotten something they can point at.
That something is the First Value Moment — the specific outcome that separates the people who stay from the people who leave. Not a step. An outcome. The distinction is the whole game: connecting your data source is a step, and seeing the first thing your data source told you is an outcome.
And the First Value Moment is not a matter of opinion. It’s sitting in your data right now, and it takes an afternoon to find:
- Pull everyone who converted and stuck for three or more billing cycles.
- Pull everyone who signed up and left inside the first 90 days.
- Look at the first 30 days — for a trial, the first session — and find the behavioral difference between the two groups.
- Whatever behavior correlates hardest with the ones who stayed is your First Value Moment.
Most teams have never run this. They have a guess, inherited from whoever built the onboarding flow, and the guess is usually a setup step because setup steps are the easiest thing to instrument.
Like this? I send one idea like it most weeks. One PLG idea a week → subscribe
The ten-user check
Before you build anything, run this. It takes about two minutes and it tells you whether you have an activation problem or a product problem, which are treated as the same thing far more often than they should be.
Pull your last ten users who churned or let the trial lapse. For each one, answer a single question: did they ever reach the First Value Moment?
If more than half of them left before reaching it, you do not have a product problem, a pricing problem, or a competitor problem. You have an activation problem, and every dollar you spend on messaging, discounts, or a better pricing page is being spent on people who never saw what you sell.
If fewer than half left before reaching it — if most of them got the value and left anyway — then your activation is fine and the problem is genuinely downstream. That’s a different project entirely.
The number of times a team has been six weeks into a pricing redesign when the honest answer was the first one is higher than anyone would like.
Build the fallback, not just the happy path
Here’s where it gets practical, and it’s the pattern I keep landing on in client sessions.
Once you accept that the goal is a real win rather than a completed setup, you have a problem: the real win usually depends on the user’s own data, and the user frequently cannot give you their own data in the first session. They don’t have the file handy. They don’t have access. They’re on a phone. The credentials belong to someone else.
A single onboarding path guarantees those people get nothing. So build value delivery as a ladder of descending friction, and let the user enter wherever they actually are:
- Their own data — the full version. Best outcome, highest friction, and the path most users can’t take on day one.
- A question with no upload — they ask about something specific in their world and get a real, specific answer back. No asset required.
- A worked example or template — for the user who stalls entirely. This one has to be concrete. A generic empty state with a “get started” button is not a fallback; it’s the same dead end with better manners.
One client I worked with recently designed exactly this: the full analysis if you upload, a product-specific answer if you’d rather just ask, and a real worked example if you do neither. The framing they settled on for the middle rung was send us what you can and we’ll give you as much value as we can with it — which is an honest trade, and converts far better than making someone feel like they failed the first screen.
Notice what this does to the gate, too. The diagnosis is free. The paywall goes where the user reaches for something actionable — the full report, the thing they’d actually send to somebody. You’re not hiding whether value exists. You’re charging for the usable form of it.
What to change on Monday
Three things, in order.
Find your First Value Moment with the cohort comparison, rather than asserting it in a meeting. Then redefine “activated” in your own analytics to mean that moment and nothing else — and expect your activation rate to drop hard when you do, because the old number was measuring setup and flattering you. Then move every trigger you own — email, in-app, sales alert — off “setup complete” and onto the real line.
The first sixty seconds matter enormously, but only because of what has to happen at the end of them. Speed to a setup screen is not the goal. Speed to something the user can point at is.
If you can’t currently name the moment that separates your converters from everyone else, that’s the first thing I do in an engagement.
Running a trial you’re not happy with? Book a call and I’ll walk your funnel live.
