Most onboarding gets designed backwards. Someone asks “what does a new user need to know?”, the team lists the features, and the list becomes a tour. It ships, it looks thorough, and the conversion rate doesn’t move.
The better question is one nobody has to guess at: what did the users who paid you actually do that the users who left never did?
That action is the Golden Path.
What it is, precisely
The Golden Path is the shortest sequence of actions that takes a new user from signup to their first real experience of value. Not “completed onboarding.” Not “reached 80% profile completion.” One concrete, instrumented behavior that separates converters from churners, predicts retention better than anything else you track, and gives every downstream decision — every email, every in-app nudge, every gate — a single test to pass: does this move the user closer to that action?
The First Value Moment is what the Golden Path delivers. The path is the sequence; the FVM is the payoff. Worth keeping separate, because you compress the path and you measure the moment.
How to find it
Start with data, not a workshop.
Pull every user who converted to paid and look at what they did in their first seven days. You’re looking for the action that shows up in roughly 80% of converters and rarely in the ones who churned. Not the most common action overall — the most discriminating one. Then sanity-check it against product intuition: the Golden Path is usually the moment the product does real work, not the moment setup finishes.
Third, confirm you can instrument it. If it isn’t a product event you can query, you can’t build a system around it, and the whole exercise quietly reverts to opinion.
The famous examples are famous because they’re unambiguous. Slack’s is a team reaching real message volume. Dropbox’s is a first file stored. Figma’s is a first frame plus a first collaborator. Same pattern every time: the product did something for the user, and the user could tell.
The trap that costs the most
Here’s where teams lose the most ground, and I hit it again with a client this month.
Their product runs on the customer’s operational data. To deliver anything genuinely useful, it needs a connection into the system where that data already lives. That connection is also, by a wide margin, the most valuable thing the company gets from any user — more data makes every recommendation across the entire system sharper.
So the instinct is obvious. Make the connection step one. It’s the most important thing, so put it first — and while we’re at it, maybe gate it behind a paid tier, since it’s so valuable.
Both instincts are wrong, and for the same reason. Connecting a data source is a high-friction, mildly nerve-wracking act that a user has no reason to perform for a product that hasn’t proven anything to them yet. Put it at minute zero and you’ve built a wall where your onboarding should be. Gate it behind payment and you’re charging people for the privilege of making your product better.
The resolution is to separate the two questions. What’s the fastest thing this product can do that a user will recognize as valuable? That’s the Golden Path, and it needs to be reachable in the first sixty seconds. The data connection comes immediately after — the obvious next step for someone who just saw a small result and now wants the real one. Same step, different position, completely different completion rate.
The general rule: a setup step is not a value moment. “Connect your account” is setup. “First successful sync” is value. If your Golden Path is a step that benefits you before it benefits the user, you’ve picked the wrong action.
Like this? I send one idea like it most weeks. One PLG idea a week → subscribe
Two other ways to get it wrong
Too broad. “Completes onboarding” isn’t a Golden Path — it’s a container holding six different behaviors with six different completion rates, and it will tell you nothing about which one mattered.
Too narrow. “Builds all five workflow templates” fails the opposite way. If a normal user can’t plausibly reach it inside the trial window, you’ve defined an aspiration, not a path.
The binary that makes everything downstream easy
Once the Golden Path is named, every trial user is in exactly one of two states.
On the path. Logging in, taking action, moving toward the behavior. Your job is acceleration: confirm the win, show the next step, get out of the way. The most common failure with on-path users is over-communicating with someone who is already succeeding.
Off the path. Signed up, stalled, hasn’t experienced anything. Your job here isn’t instruction, it’s intent. This user doesn’t need a better tutorial; they need a reason to come back. Lead with what someone like them got out of it, not with a numbered list of steps.
That binary is most of the reason to do this work at all. Before writing any email or in-app message you answer one question — which state is this user in — and the content decision is nearly made for you. It’s also the spine of the 4-Profile System: the profiles are just positions relative to the path.
Time to first value is the lever
Every day between signup and the First Value Moment is a day the user can leave, and most of them do it quietly. That’s why compressing time-to-FVM outranks nearly every other activation project.
Two numbers from recent client work make the point from opposite ends. At one infrastructure client, roughly 10 of every 250 signups were reaching genuine value — the other 240 weren’t churning after a bad experience, they were churning before having one. At another, the first paying customer converted in about six minutes with no hand-holding at all. Same lever, opposite readings: one product’s path was too long to survive, the other’s was short enough that the guided tour had become the obstacle. That second one carries its own lesson — build an exit from onboarding for the users who don’t need it.
Neither number is a benchmark to borrow. They’re both just evidence that the distance between signup and value is a design decision, not a fact of nature.
Where this sits in the system
The Golden Path is the first thing I identify in an engagement, and it feeds everything after it. You can’t place a value gate correctly until you know where value happens. You can’t score product-qualified leads without knowing what qualified behavior looks like. You can’t segment a lifecycle sequence by anything that matters. It’s the middle layer of the 10 Laws of Conversion for exactly this reason — the laws above it define the machine, and the laws below it all consume its output.
There’s a discovery question I use that sorts companies fast: what’s the one action inside your product that, once a trial user does it, they almost never leave?
If you can answer that with data, you’re further along than most. If you can’t — and most can’t — that isn’t a gap in the interview. That’s the first piece of work, and it’s worth doing on its own before you touch a single email.
Running a trial you’re not happy with? Book a call and I’ll walk your funnel live.
