Performance Evaluation Checklist

Pull the eligibility roster from the HRIS and apply the cycle rules — typical exclusions are employees in their probationary window, those on FMLA or PFML during the review period, and recent transfers under 90 days. Flag any borderline cases for the HRBP before locking the list.

Export OKRs or SMART goals from the performance system along with any documented 1:1 notes, project retros, and prior PIP records. The narrative is much harder to write defensibly without the supporting record from the cycle.

Compensation review riding on this cycle requires an accurate exempt/non-exempt classification and current band/midpoint data. Re-confirm the duties test for anyone whose role changed mid-cycle, and pull market data from Pave, Radford, or PayScale before calibration.

Launch the self-review with a clear due date and a note on what reviewers will and will not see. Self-reviews submitted under time pressure on the last day are a common quality issue — give a real two-week window.

Solicit three to five peer reviewers per employee, mixing cross-functional partners and direct collaborators. Avoid letting the employee hand-pick only friendly reviewers — the manager and HRBP should sign off on the list before it goes out.

Tie each rating claim to a documented example — project shipped, quota attained, customer escalation handled. Vague narratives ("strong contributor") collapse under scrutiny if the rating later supports a separation or PIP. Write to the SBI model: Situation, Behavior, Impact.

Use the published level-and-rubric — not gut feel. Common bias to watch: recency (last sprint dominates the cycle), halo (one strong project lifts every competency), and leniency (everyone meets expectations). Calibration will catch some of this; the rubric prevents most of it.

Drafts must be locked before the calibration session — late submissions push the whole cohort review. The HRBP runs the room; managers defend ratings against peer ratings on the same level so the rubric is applied consistently across the function.

Update the rating, the narrative, and the supporting examples — a rating that moved in calibration without an updated narrative is a red flag in any later employment dispute. The narrative must read as if the calibrated rating was the manager's own conclusion.

Scan for protected-class language (age, pregnancy, disability, accent, religion), gendered descriptors ("abrasive," "too aggressive," "emotional"), and references to FMLA / PFML / accommodation usage. These are the patterns that show up in Title VII, ADA, and PWFA claims sourced from review documents.

Book a private 45–60 minute room or video call — never deliver a Below Expectations rating in a public space or back-to-back with another meeting. Send the written review to the employee 24 hours before the conversation so they can read it without the manager watching.

Capture the employee acknowledgment signature in the performance system. Acknowledgment confirms receipt — it is not agreement with the content. If the employee declines to sign, document the refusal and have the HRBP countersign as witness.

Goals carry forward into the next cycle's review — make them measurable and dated. Avoid "improve communication" as a standalone goal; convert it to an observable behavior with a cadence ("deliver weekly status update by EOD Friday").

A PIP issued the day before separation reads as pretext in court. Build genuine 30/60/90 milestones with measurable deliverables, weekly coaching cadence, and documented manager check-ins. Loop employment counsel before delivery if the employee is in a protected class, on leave, or recently filed a complaint.