The Same Briefwhere T and π diverge in practice

The onboarding flow for a consumer fintech app has a drop-off problem. Analytics show that a large share of new users do not complete step 3 of a five-step sign-up. The product manager briefs the designer: Reduce drop-off at step 3. We have three weeks.

Both designers, the T-shaped and the π-shaped, receive the same brief. Both are senior, with enough breadth to read a flow, talk to engineering, and understand why onboarding matters commercially. Both will diagnose before they design. The difference shows in what happens after the diagnosis.

The T-shaped designer has an upstream vertical, so they do not take the brief as handed. They pull the analytics more carefully, looking past the drop-off rate to the drop-off timing. Users are abandoning after spending longer than average on the screen, with no immediate bounce. Something is making them wait and then leave. They talk to engineering. Step 3 triggers an email verification from a legacy auth system, and the email takes forty to ninety seconds to arrive. Users are leaving because they are waiting, and nothing in the flow tells them to wait. The drop-off looks like a UX problem. It is a backend constraint.

That diagnosis produces a different brief: bridge the wait in place of redesign step 3. They design a holding state that signals the verification is in progress, and a session-recovery flow for users who leave and return. Three weeks later, drop-off at step 3 has fallen sharply.

The π-shaped designer reaches the same diagnosis by the same route and ships the same fix. Then the second vertical goes to work. They write a short memo to the PM making the case for replacing the legacy auth system in the next planning cycle, a strategic document that makes a technical constraint legible as a business priority. They walk the team through how the drop-off was misread, so the next brief that looks like a screen problem gets a timing check first. Six months later the auth replacement is on the roadmap, and another team has used the same check to find a hidden wait of its own.