Offboarding, Backfill & Closing the Loop

The last part closes the system. It covers the mechanics of a clean exit, how backfill planning folds back into the hiring half, and the process-hygiene rule that keeps exits from stalling in limbo. This is where the firing half hands back to the hiring half, and the loop becomes a loop.

An exit cannot be processed on a hallway comment. Whether someone resigns or is terminated, the trigger must be in writing — an email with the last day is enough — and people ops must be notified immediately. A person who says "I think I'm going to leave" has not resigned; until there is a written record with a date, the exit stalls and no downstream step can start. The same discipline applies in both directions: written trigger, notify people ops, record in the system of record.

People ops typically owns most of this once the termination is entered in the HRIS; the manager owns knowledge transfer and team communication. Run all four tracks:

Access & security — revoke single-sign-on on the last day so all downstream apps close; remove privileged group, cloud, repository, secret, on-call, and admin access; rotate any shared secrets or tokens the person held; reclaim equipment.

Data & continuity — forward the departing person's email to the manager after the exit date; transfer their document drive to the manager; reassign owned tickets, docs, code-owner entries, and dashboards.

Knowledge transfer — hand off in-flight work, runbooks, and tribal knowledge while access is still live during the notice window; explicitly document anything only that person knew, such as accounts they owned or systems no one else touches.

Payroll, benefits & records — process final pay (or the final invoice for a contractor), terminate benefits, update the HRIS effective the last day, and hold an exit interview for employees.

Then handle communication: tell the team factually and respectfully, without disparaging the person, redirect their in-flight responsibilities, and update the org chart and reporting lines.

Do not treat backfill as an afterthought. Decide it inside the exit decision so recruiting can move immediately. Recompute against the floor — the minimum viable engineer count per team or pillar — because several concurrent exits can drop you below it fast; quantify it plainly ("after these exits we sit at N, below the floor"). Then apply the headcount discipline from Part 05: separate approved headcount from realistic hiring throughput, sequence direct backfills before strategic adds, decide the role mix before opening reqs, and leave slack for one or two more departures in the planning window.

Notice what just happened: the backfill re-enters the hiring loop, at the level the exit revealed the team needs, screened against the same pillars, onboarded against the same principles, and reviewed against the same yardstick. The exit did not just remove a person — it fed information back into the front of the system about where the bar needs to hold and what the team is actually missing.

That is why hiring and firing belong in one system rather than two. The pillars that gate the interview are the pillars a new hire is measured on from day one and the pillars an underperformance case is built against. The dignity that governs an offer — being honest about level, closing in good faith — is the dignity that governs an exit. Run the two halves against one written bar and the whole thing compounds: each hire sharpens the bar, each exit sharpens the backfill, and the team you end up with is the team the system selected for, on purpose, out loud, and on the record.