Most people quietly think performance management is broken. Is it — or is it just being done badly?

The frustration is real and widespread — and so is the reported payoff where it's done well. A McKinsey Global Survey of managers and executives (summarised below in our own words — see Sources; we don't reproduce McKinsey's text or exhibits) found that more than half of respondents said their organisation's current system has no positive effect, or a negative one, on either employee or company performance, and two-thirds had already changed something in the previous 18 months without consensus on what to fix. Yet among organisations whose systems were rated effective, 60% reported having outperformed peers over the previous three years — nearly three times the share with ineffective systems. Read that as an association, not a causal effect: respondents rated their own system and their own relative performance, so the two judgements are not independent. The useful signal is that dissatisfaction concentrates on one specific dimension.

That dimension is perceived fairness — and this is where the independent evidence is strongest. In the McKinsey survey, whether people experienced the system as fair was the factor most strongly associated with reported effectiveness (60% of those who saw it as fair called it effective overall, versus 7% who didn't) — an association within a single self-reported survey, not evidence that fairness alone drives company performance. That association isn't a McKinsey idiosyncrasy: a 25-year meta-analytic review of organisational-justice research (Cohen-Charash & Spector, 2001 — 190 study samples, 64,757 participants) found that of the justice dimensions, procedural justice — the perceived fairness of the process — is the one most strongly tied to job performance and to counterproductive behaviour. Fairness of process, not the rating curve, is the hinge. Three manager practices were associated with stronger reported effectiveness.

Practice 1 — link goals to priorities, and revisit them. In the McKinsey data, where individual goals were tied to strategy, 46% reported effective performance management versus 16% where they weren't, and among effective systems 62% revisit goals at least twice a year. The underlying goal-setting science supports this: specific, concrete, appropriately challenging goals are reliable motivators that lift performance more than vague ones (Höchli, Brügger & Messner, 2018, Frontiers in Psychology; Höpfner & Keith, 2021, Frontiers in Psychology). But the same research carries a warning that maps onto the "revisit" point — over-narrow goals can produce tunnel vision, and missing a fixed goal damages motivation and effort — which is exactly why goals that are revisited as conditions change stay fair and useful rather than punitive.

Practice 2 — coach, don't just rate. In the McKinsey survey, effective coaching was the practice most strongly associated with perceived fairness, yet fewer than 30% of respondents said their managers actually do it; where managers coached effectively, 74% called the system effective, versus 15% where they did not. Crucially, coaching behaviour is trainable, not a fixed trait: a randomised, waitlist-controlled trial of managerial behavioural training (Grill, 2025, PLoS ONE; 49 managers and 439 employees in one Swedish municipality) found employees whose managers were trained reported significantly greater improvement in their managers' performance-feedback behaviour (β = 0.30, p = 0.04), sustained across the full 18-month follow-up. Note the outcome that trial actually measured: managers' feedback behaviour as rated by their employees — not company performance. Manager coaching capability, the thing most often missing, is the thing you can most directly build.

Practice 3 — differentiate pay by performance. Fewer than half of McKinsey's respondents said pay meaningfully varied by performance level; where it did, 54% rated the system effective versus 16% where it didn't, and separating pay conversations from the review itself was associated with higher reported effectiveness again (47% vs 30%). Organisations strong on all three practices were 12 times likelier to report an effective system than those doing none. Technology tracked only a marginal difference; the fairness practices carried the association.

The real split isn't "keep ratings vs kill ratings." It's whether you treat performance management as a form to optimise or as a fairness problem to solve — the second is where the reported outperformance clusters.

Don't redesign the form — build the fairness. Performance management isn't failing because the rating scale is wrong; it's rated worst where people don't believe the process is fair, and perceived fairness travels with three concrete manager behaviours rather than with a better template.

• What to rely on: goals that visibly connect to strategy and get revisited; managers who actually coach and give ongoing feedback — a capability you can train; and pay that meaningfully reflects performance. Organisations strong on all three reported effective systems far more often than those doing none.
• What to avoid: treating a software upgrade or a shorter form as the fix; one-and-done annual goals and over-narrow high-stakes targets; and "differentiation" that exists on paper but not in anyone's paycheck.
• The point that matters: the most promising lever is the least glamorous — manager capability. Fewer than a third of managers coach well, coaching is the practice most strongly associated with perceived fairness, and a randomised trial shows manager feedback behaviour can be improved by training. Start there.

1. Ask the fairness question directly. In your next skip-level or engagement pulse, ask whether people believe the performance process is fair — in this survey that perception tracked reported effectiveness more closely than any mechanical feature of the system did.
2. Trace three goals to the strategy. Pick three employees and check whether their current goals visibly connect to a business priority — and when they were last revisited. If you can't draw the line, neither can they.
3. Audit manager coaching, not just manager ratings. Find out how many of your managers hold regular development conversations. If it's under a third, that's your real bottleneck — and it's trainable.
4. Check whether pay actually differentiates. Compare rewards for your strongest and weakest performers. If they're barely different, your "high performers" are learning that performance doesn't pay.
5. Separate the pay talk from the review. Splitting the "how did you do" conversation from the "here's your number" conversation is a small change that respondents who made it rated more effective (47% vs 30%).

This section is Peoplense inference, labelled as such. None of the evidence above is Gulf-specific — the McKinsey survey, the justice meta-analysis, the goal-setting research and the Swedish training trial are all international. What follows is our reading of how that evidence may apply here. It is not a finding about Gulf workplaces, and we have not tested it against Gulf data.

Gulf organisations working to Vision-2030 targets already speak the language of measurement — KPIs, targets and benchmarking are common — which we think helps: a culture comfortable with "show me the number" has part of a disciplined performance system in place already. But measurement is not the same as fairness, and rigid, high-stakes targets are precisely what the goal-setting research warns about. Where organisations are scaling quickly, leaning on newer or first-time managers, or running large mixed-nationality workforces, we would expect the two factors above — whether people experience the process as fair, and whether managers coach rather than merely score — to matter more rather than less. On that reading, the move for Gulf leaders isn't more sophisticated forms; it's building manager capability so the measurement culture the region already has is also felt as fair. We would replace this inference with evidence if Gulf-specific research emerges.

Honest scope: the integrating survey here is McKinsey's, summarised with permission — self-reported and correlational, not causal, and respondents rated both their own system and their own relative performance, so those two judgements are not independent of each other. Source independence: McKinsey sells performance-management and people-and-organisation consulting services, so it has a commercial interest in the finding that these systems can be made to work. We record that as a factor in assessing the source, not as grounds to dismiss it. The justice meta-analysis and the goal-setting studies are international and not performance-management-specific; the randomised trial covers manager training and feedback behaviour in one Swedish municipality, not firm performance; and none is Gulf-specific. The Gulf read-across is our argued interpretation, labelled as such above. Treat the percentages as signal, not settled fact.

Featured source — summarised with McKinsey's written permission (original, own-words summary; no McKinsey text quoted and no exhibits reproduced; McKinsey named as the source with a direct link, per the permission granted 2026-07-10):

• McKinsey & Company, Harnessing the power of performance management (McKinsey Global Survey) — original. Reported effectiveness is most strongly associated with perceived fairness; three practices track it — linking goals to priorities, manager coaching, and pay differentiation — and organisations strong on all three are 12x likelier to report an effective system. Self-reported and correlational. Source-independence note: McKinsey sells people-and-organisation and performance-management consulting services, a commercial interest we record here as part of assessing the source.

Library / open-licensed sources (Creative Commons; the peer-reviewed evidence behind the mechanisms):

• Grill, M. (2025), Recognizing employees' contribution to effectiveness and values: A randomized waitlist-controlled trial of operant-based leadership training, PLoS ONE, 20(4):e0320131 — original · licence: CC BY 4.0. Randomised, waitlist-controlled trial in one Swedish municipality (49 managers, 439 employees): training significantly improved employees' ratings of their managers' performance-feedback behaviour (β = 0.30, p = 0.04), sustained over 18 months. The measured outcome is manager feedback behaviour, not organisational performance — it supports the claim that coaching capability is trainable, and no more than that.
• Höchli, B., Brügger, A. & Messner, C. (2018), How Focusing on Superordinate Goals Motivates Broad, Long-Term Goal Pursuit: A Theoretical Perspective, Frontiers in Psychology, 9:1879 — original · licence: CC BY. Challenging, specific, concrete goals are powerful motivators that boost performance more than vague ones — while narrow subordinate goals risk an overly narrow focus of attention.
• Höpfner, J. & Keith, N. (2021), Goal Missed, Self Hit: Goal-Setting, Goal-Failure, and Their Affective, Motivational, and Behavioral Consequences, Frontiers in Psychology, 12:704790 — original · licence: CC BY. Specific, high goals are among the best-established tools for lifting performance and motivation — but failing a goal carries real affective and motivational costs, a caution against rigid, high-stakes targets.

Cited findings (named and linked, not republished — these do not carry an open licence):

• Cohen-Charash, Y. & Spector, P. E. (2001), The Role of Justice in Organizations: A Meta-Analysis, Organizational Behavior and Human Decision Processes, 86(2), 278–321 — publisher. Across 190 study samples (64,757 participants), procedural justice was the justice dimension most strongly related to job performance and counterproductive work behaviour. Cite-only.
• Locke, E. A. & Latham, G. P. (2002), Building a Practically Useful Theory of Goal Setting and Task Motivation, American Psychologist, 57, 705–717 — ERIC EJ654871. The foundational synthesis behind "specific and challenging goals outperform vague goals." Cite-only.

A related Peoplense brief that puts the same "is this practice actually working?" lens on performance:

• Should we remove performance ratings? — the ratings-free debate, measured.