Why does B2B onboarding fail even when the UX is good? Because the flow is designed for 1 person and 3 different people show up. The buyer needs evidence the decision was right, a number in week 1. The admin needs confidence nothing breaks: permissions that map to the org, an integration that connects cleanly. The end user needs their actual task done faster today. A first session that serves 1 of them taxes the other 2, and the analytics report an average of 3 experiences, which describes none of them.
The fix is to branch. Detect the persona from invite type, role, SSO metadata and the first 3 clicks, then define activation per persona and instrument the chain between them. The account is activated when all 3 links hold, not when any 1 person succeeds.
B2B onboarding keeps failing for a reason no amount of UX polish fixes: the flow is designed for 1 person, and 3 different people show up. The buyer who signed the contract, the admin who has to wire the product into everything, and the end user whose week it is supposed to improve. Each arrives with a different definition of value, and a first session that serves 1 of them taxes the other 2.
Who actually shows up?
The buyer needs evidence the decision was right: a number, a report, proof in week 1 that the business case is materializing. The admin needs confidence nothing breaks: permissions that map to the org, an integration that connects cleanly, a test run that behaves. The end user needs their actual task done faster, today, without learning a philosophy first. A single onboarding flow optimizes for whichever of the 3 the designer imagined, and the analytics then report an average of 3 different experiences, which describes none of them.
The split is not only an onboarding problem. It is the same committee that decides an enterprise purchase, which is why account-level qualification and account-level onboarding are the same discipline seen from 2 ends.
How do you detect which persona showed up?
The persona is knowable at the front door. Invite type, role field, SSO metadata, and the first 3 clicks all speak: an admin heads for settings and integrations, a user heads for the work, a buyer heads for the dashboard. Ask 1 honest question at entry if the signals disagree. Then branch the first session, because the cost of branching is 3 flows to maintain and the cost of not branching is 2 personas bouncing off a flow built for the third.
What does activation mean when 3 people have to succeed?
The account is not activated when any 1 persona succeeds. User habit without admin configuration dies at rollout. Admin configuration without user habit is shelfware with clean permissions. Buyer proof without either is a renewal argument you will lose in 11 months. Define activation per persona, instrument the chain between them, and find which link breaks in which segment, because that link, not the average funnel, is the roadmap.
The renewal is won in week 1
Expansion and renewal decisions trace back further than anyone budgets for: to whether the buyer saw their proof while the users were still building the habit and the admin still remembered the setup going smoothly. That is retention work happening long before anyone calls it retention. The personas were always plural. I made the same argument about humans and agents in When Your Next User Is an Agent, and the discipline is identical: define activation for each evaluator you actually have, then measure each one honestly.
Onboard the account, not the login
Design 3 first sessions. Define 3 activation moments. Instrument the chain that connects them. Renewal gets decided by the account's experience, and the account is 3 people with 3 different reasons to stay.