Writing a Review for a Direct Report You Inherited

You have three months of direct observation covering a twelve month review period, and the rest of the file is other people's notes, a predecessor's Slack messages, and whatever the employee tells you themselves. That gap is not an excuse to write a thin review. It is a specific problem with specific fixes.

The principles

These aren't steps to complete in order - they're the things worth checking yourself against, before and during the work. Read them once, then keep this page open while you go.

1. Get the predecessor's actual notes, not just their handover summary

   A handover conversation is a manager's summary of a summary, filtered through whatever they remembered on the day they left. The original 1:1 notes, past self assessments, and any prior ratings are closer to the source and catch things a verbal handover conveniently smooths over.

   • Ask HR or the outgoing manager directly for 1:1 notes, prior review documents, and any goal-tracking artifacts, not just a verbal summary.
   • If the predecessor is still reachable, ask them one specific question in writing: "What would you flag as this person's strongest and weakest area this period?" A written answer holds up better than a hallway comment you're paraphrasing from memory.
   • Check the dates on everything you're given. A note from month two of the period is worth less to a rating that covers months nine through twelve, and you need to know which is which.

2. Collect peer and cross-functional input deliberately, not by accident

   When you didn't see the work happen, the people who did are your actual evidence base. Waiting for this input to surface on its own means you'll only hear about what was memorable, which skews toward recent and toward extremes.

   • Identify three to five people who worked closely with this person for most of the period you didn't observe, not just whoever is easiest to reach.
   • Ask a specific question, not a general one: "What did they own, and how did it go?" beats "any feedback on so-and-so?" by a wide margin in what comes back.
   • Push for named, dated examples rather than overall impressions. "They're solid" tells you nothing you can put in a review; "they redesigned the onboarding flow in March and it cut support tickets" does.
   • If two sources contradict each other, do not average them into something noncommittal. Go back to one of them and ask what specifically they mean.

3. Weight the period honestly, not evenly and not by recency

   The instinct is either to treat the whole year as one blob you now understand equally well, or to quietly let your own three months of direct observation dominate because it's what you actually trust. Both are distortions, just in opposite directions.

   • Write down, honestly, what fraction of the period you directly observed versus reconstructed, and let that fraction show in how much weight each part gets in your reasoning, not just in your head.
   • Do not let a strong or weak few weeks under you outweigh ten months of documented work under someone else. That's recency bias wearing an inherited-manager costume.
   • If the earlier period and the period you've observed tell genuinely different stories - say, a strong first three quarters and a shaky one under you, or the reverse - name that shift explicitly rather than blending it into one flat number.

4. Ask the employee to fill the gap themselves

   A self assessment is useful for anyone, but it is disproportionately valuable here, because the person you're rating has ten or eleven months of context that currently exists nowhere else in your evidence.

   • Ask for a self assessment structured around the same criteria you're rating against, with specific examples and outcomes, not a general narrative of the year.
   • Treat gaps between what they report and what you can independently verify as things to ask about directly, not as automatic red flags. A manager transition is exactly the kind of period where documentation naturally gets thin.
   • Ask them directly what they think their previous manager would have said about their year. It is a strange question, but it usually surfaces the exact thing they're worried didn't make it into the handover.

5. Write the review so it shows its own sourcing

   A review that reads as confidently as one where you watched every week of the work invites a fair question: how would you know that? Naming your source per claim is what makes a review built on secondhand evidence still trustworthy.

   • Where a point of evidence comes from a peer, the predecessor, or the employee's own self assessment rather than your direct observation, say so in a sentence, not as a caveat buried in an appendix.
   • Avoid writing evaluative language you can't back with a source. "Consistently reliable all year" is not something you can claim about ten months you weren't there for, even if it's probably true.
   • Keep the parts you did observe directly written with full confidence. The goal isn't to hedge the whole document, just to be accurate about which parts are which.
   • Once you've pulled the evidence together, a tool like Perform Review's AI Performance Reviews can help turn it into a first draft, which is often easier to edit for accuracy than to write from a blank page when half your sources are secondhand.

6. Flag the confidence gap to whoever calibrates the rating

   If this review feeds into a calibration process, the room needs to know this rating rests on a shorter observation window than everyone else's, or your read gets treated as equally solid when it isn't. Either an inflated or deflated number can slip through unchallenged as a result.

   • Say plainly, in whatever notes accompany the rating, how much of the period you personally observed.
   • If you're genuinely unsure between two adjacent ratings because of the evidence gap rather than the performance itself, say that too. It's a different problem than an unsure rating based on ambiguous performance, and it deserves a different response.
   • Ask your own manager or HR whether there's a standard practice for reviews spanning a manager transition. Some companies have one and don't advertise it. It's worth one email to check before you improvise something.

7. Tell the employee how the review was built, before they read it

   Someone who spent most of the year under a different manager and is now being rated by someone who barely knows them is primed to distrust the number, fairly. Naming your process up front, before they see the rating, defuses far more of that than defending it after the fact would.

   • Open the review conversation by explaining briefly how you built it: predecessor's notes, peer input, their own self assessment, and your own direct observation.
   • Invite correction specifically, not generally. "Is there anything from before I started that you think I'm missing" gets a much more useful answer than "any questions."
   • If something in the review turns out to be wrong because a source was incomplete or outdated, fix it and say so plainly. Getting caught trying to quietly patch it costs you more trust than the original gap ever would.

That's the whole playbook

None of this replaces judgement. It is the checklist for catching the moments where judgement quietly turns into habit, or into bias.