Persona Based Onboarding: Different First Runs for Different Users
Onboarding is where personas stop being research posters and start costing or making money. A first time user trying to solve an urgent problem and a first time user exploring on behalf of a team need different first runs. One generic flow, designed to show everyone everything, usually teaches each of them that the product is not quite for them.
Branch on the job, not the demographics
The useful onboarding question is not who the user is but what they came to do first. Two or three job based branches, chosen from your primary personas, cover most products: the urgent solver, the evaluator comparing options, and the team organizer setting up for others. Each branch should reach a first success in minutes, by a route matched to that job.
Ask the branch question plainly at signup, with the personas own language for the options, and always allow the generic path for users who skip. Forced classification with no exit produces garbage answers and annoyed users. A skipped question is still data; a resentful one is noise with a grudge.
What actually changes between branches
Order, emphasis, and defaults: which feature is introduced first, which example data illustrates it, which templates are offered, and which invitations are suggested. The urgent solver meets the core action and a realistic sample of their own problem. The evaluator meets comparison friendly views, export options, and the evidence they will need to justify a choice internally.
Keep the shared core shared. Branches that diverge into separate products become a maintenance trap within a year. The healthy pattern is one product, a few ordered paths through it, and defaults that differ. If a branch needs its own feature set to work, your personas may be describing two products wearing one name.
Measure branches against their own promise
Judge each branch on its personas success metric: time to first completed job for the urgent solver, evaluation artifacts created for the comparer, teammates invited for the organizer. A single blended activation rate hides the branch that is failing behind the branch that is coasting.
Watch branch selection data too. If most users pick one branch, your persona split may be wrong or the question may be leading. If branch performance converges, the branches have stopped mattering and should be simplified back toward one flow with good defaults.
Onboarding is a persona test you run daily
Every new user who misclassifies, churns instantly, or succeeds by an unplanned route is evidence about your persona set. Feed onboarding outcomes back into persona reviews: which archetype never appears, which appears constantly but has no branch, and where users consistently describe their job in words your personas do not use.
That feedback loop is the real prize. Research personas are hypotheses. Onboarding is the one place the market votes on them every day, in behavior, at scale, for free. Products that listen end up with fewer, truer personas and first runs that keep getting shorter.
Keep reading
Related guides on this site: Negative Personas: Deciding Who You Are Not Building For Running a Persona Workshop That Ends With Decisions, Not Posters Practical answers about personalization.
Frequently asked questions
How many onboarding branches should a product have?
Two or three, matched to your primary personas jobs. More branches than that usually means the persona set is too fragmented, and each additional branch multiplies the testing and maintenance the flow needs.
What if users pick the wrong branch?
Make switching free and visible, and let behavior override the initial pick: a user who chose the evaluator path but starts doing solver work should see solver defaults emerge. The initial answer is a starting point, not a sentence.
Should onboarding ask about job or role?
Job first. Role titles vary wildly between companies, while the task someone came to complete is concrete and predictive. Role can refine defaults later, once the product knows the user better.
How do we know a branch is failing?
Compare it against its own personas success metric, not the blended rate. Slow time to first success, high early abandonment, and users switching branches mid flow are the clearest signals.