When a senior hire fails, audit the onboarding before you blame the hire
Month-three failures usually trace to four org-side gaps, not to the person you hired
When a senior hire is struggling at month three, the postmortem almost always points at the person. Wrong seniority. Bad culture fit. Interview process missed something. Sometimes that is true. But having been the new senior more than once across 16 years, and having watched other seniors land well or badly, I think the org fails the hire more often than the hire fails the org.
Four failures show up again and again, and none of them appear in the hiring postmortem.
Nobody gave them the map. Every codebase has a real shape that lives in people's heads: the service that pages at night, the module everyone routes around, the migration nobody wants to touch. A senior's whole value is making good decisions against that map. If acquiring it takes months of archaeology because nothing is written down and nobody made introductions, you are paying senior rates for spelunking.
The docs lie. Worse than missing documentation is documentation that is confidently wrong: the README with setup steps that stopped working two quarters ago, the architecture diagram from before the rewrite. A new senior burns their credibility-building weeks discovering which docs to distrust. In my own monorepo I ended up making doc updates part of the definition of done, enforced in review, precisely because stale docs tax every future person silently.
Decision authority is undefined. You hired them to make judgment calls, then left it unstated which calls are theirs. So they either overstep and get burned, or under-step and look passive. Both read as "not working out" by month three. The fix costs one conversation in week one: here is what you decide alone, here is what you bring to me, here is what the team votes on.
Their first tickets are trivia. Starter tickets are supposed to be safe, but three weeks of typo fixes teaches a senior nothing about the system's real risks and teaches the team nothing about the senior. Give them one real, bounded problem early. The point of week four is a real change through the full pipeline, with the machinery respected.
The pattern across all four: seniors do not fail because they cannot code. They fail because the context they need to exercise judgment was never transferred, and judgment without context looks exactly like incompetence.
So before the month-three conversation about whether the hire was a mistake, run the cheaper audit first. Did they get the map, honest docs, explicit authority, and a real problem? If the answer is no four times, the interview process was never the bug.