Performance management health check

The complete guide to the performance management health check

The health check above takes about two minutes. This guide explains why the four areas it measures matter, how to answer the eight questions honestly, how to read your score and your biggest gap, a worked example, the mistakes people make when they take it, and what to do next. It is written for CPOs, Heads of People and founders at growing companies who want a fast, structured read on where their performance process is weakest before they spend time or money fixing it.

Why a health check comes before a fix

Performance management problems tend to show up as symptoms: managers complaining about the time reviews take, employees saying ratings feel political, leaders unable to say who their top performers are. The temptation is to treat the loudest symptom, often by buying a new tool or redesigning the form. But different causes produce similar symptoms. Ratings that feel unfair might come from missing standards, from no calibration, or from managers working from memory. Each needs a different fix.

A structured check forces the diagnosis first. McKinsey’s 2024 article “In the spotlight: Performance management that puts people first” makes the same point: companies should start by understanding where they are now, including how well the existing system is adopted and how good it is, before deciding what to change. The same article notes that leaders often lack confidence in their performance systems, citing fragmentation, informal “shadow” systems, misalignment and inconsistency. The four areas in this health check map closely onto those complaints.

The CIPD’s performance management factsheet also frames performance management as a joined-up system (objectives, feedback, development and review working together) rather than a single annual event. A health check is useful precisely because it looks across the whole system at once, rather than at the review form in isolation.

The four areas, and why each one matters

The areas are connected. Weak standards make bias worse, because without a written bar each manager uses their own. A heavy admin burden makes bias worse too, because managers who are short of time fall back on what they remember from the last few weeks. That is why the check reports a biggest gap: fixing the weakest area usually relieves pressure on the others.

What a healthy process looks like

It helps to picture the top answer in each area, because that is what the questions are measuring against. In a healthy process, every role and level has a written scorecard that managers actually use in reviews, and one shared talent bar means a rating in Sales carries the same weight as a rating in Product. Goals live in one place, link visibly to team and company priorities, and get updated through the cycle rather than rediscovered at review time. Ratings go through a structured calibration with evidence and distribution checks before anyone sees them, and reviews draw on goals, feedback and notes from the whole period. Review prep takes minutes rather than hours because the evidence is already collected, and the whole cycle runs in one system.

Very few growing companies hit the top answer on all eight questions, and they do not need to. The point of the picture is direction: each step up on any question removes a specific failure, such as a rating nobody can defend, a goal nobody owns, or a manager who spends a weekend on review write-ups.

How to take the health check

There are eight questions, two per area, each with four answers scored from 0 to 3. A few tips so the result is useful:

• Answer for the typical case, not the best team. If one department has excellent scorecards and the rest have none, the honest answer is “written for some roles, used unevenly”.
• Answer for what happened last cycle, not what the process document says should happen.
• Ask a second person to take it. A People lead and a line manager often score the same process very differently. The gap between their answers is itself a finding.
• Do not overthink it. The check is a two-minute triage. Your first instinct about which answer fits is usually right.

Your answers stay in your browser, and the tool never asks about named employees, so you can answer candidly.

How to interpret your score

You get three things back: an overall score out of 100, a score for each of the four areas, and your biggest gap with a recommended first fix and a related customer story.

These bands are the tool’s own thresholds. The score is a checklist score from your answers, not a benchmark against other companies, so the most useful comparison is with your own score after the next cycle.

The area scores matter more than the total. Two companies can both score 50 for very different reasons: one with strong standards and chaotic admin, another with slick tooling and no shared bar. The recommended fix follows the biggest gap: write scorecards for standards, link goals for alignment, calibrate with evidence for bias, or consolidate evidence in one system for admin burden. Each gap links to a published customer story from Cleo, Plum, Ebury or Sertis that shows how one company tackled it.

Worked example: a 300-person fintech

An illustrative company (invented for the example) answers as follows:

• Standards: role descriptions exist for some roles (1); teams probably do not share a bar (1). Area: 2 of 6.
• Alignment: goals live in a tool but are rarely updated (1); most employees could name the priority they support (2). Area: 3 of 6.
• Bias: an opinion-led calibration meeting (2); reviews draw on memory plus a few notes (1). Area: 3 of 6.
• Admin: about an hour of prep per report (2); a cycle touches two systems (2). Area: 4 of 6.

Total: 12 of 24 points, which the tool shows as 50 out of 100, “Solid, with clear gaps”, with Standards flagged as the biggest gap. The People lead’s instinct had been to buy a calibration tool, because calibration meetings felt painful. The check suggests the pain is upstream: calibration is opinion-led because there is no written bar to calibrate against. The first fix is to write a talent bar by seniority and scorecards for two critical job families, run one cycle with them, then revisit calibration. Re-taking the check after that cycle shows whether the standards score moved and whether the bias score followed.

Common mistakes

• Answering aspirationally. Scoring the process you are about to launch rather than the one people experienced produces a flattering number and the wrong priority.
• Fixing everything at once. Four simultaneous changes make it impossible to tell what worked, and managers resist a process that changes every quarter.
• Treating the score as a benchmark. It does not compare you with other companies. Use it to track yourself over time.
• Ignoring the second-lowest area. If two areas are close, check which one causes the other. Standards usually come before bias, and evidence collection before admin speed.
• Skipping the conversation. The score is most useful as a prompt for a 30-minute discussion with your leadership team about which gap to close first.

What to do next

Take the result to the matching resource. For standards, use the performance management framework and the rating scales guide. For alignment, read the guide to aligning OKRs with performance reviews. For bias, run the bias checker and the guide to reducing manager rating bias. For admin burden, size the problem with the review cost calculator and read switching from spreadsheets to performance management software.

Sources: McKinsey, In the spotlight: Performance management that puts people first (2024); CIPD, Performance management factsheet; Buckingham & Goodall, Reinventing Performance Management, HBR (2015).