Performance management (PM) is the continuous process of setting expectations, tracking progress, giving feedback, developing people and rewarding results, so that individual effort supports organisational goals. Performance appraisal is just one step of it, the formal assessment. Treating the yearly review as the whole system is a common mistake.
1. The performance management cycle
| Stage | Activity |
|---|---|
| Plan | Agree goals, standards and development needs |
| Monitor and support | Track progress, coach, remove obstacles |
| Review | Assess results and behaviours; discuss |
| Reward and develop | Link to pay, promotion and learning; set the next cycle |
Performance has three components: results (what), behaviours (how) and development (growth). A good system is fair, transparent, clearly linked to strategy, regular, and developmental, not only judgemental. Purposes: administrative (pay, promotion, termination) and developmental (feedback, training); mixing them can reduce honesty.
2. Setting goals and measures
| Term | Meaning | Example |
|---|---|---|
| Key result area (KRA) | The area of responsibility | Delivery quality, customer service |
| Key performance indicator (KPI) | A measure of performance | On-time delivery %, defects per release |
| Target | The level expected | 98% on-time |
| Weighting | Importance of each goal | 40% / 30% / 30% |
SMART goals: Specific, Measurable, Achievable, Relevant, Time-bound. Locke’s goal-setting theory: specific, challenging, accepted goals with feedback lift performance more than “do your best”. Management by objectives (MBO) (Drucker) has managers and subordinates set goals together and review them.
OKRs (Objectives and Key Results): an objective is a qualitative, inspiring aim; key results are measurable outcomes (typically 3–5). They are often set quarterly, shared openly and graded; stretch goals are expected (hitting 70% may count as success).
Example: Objective: “Make onboarding delightful.” Key results: reduce time to first order from 7 days to 3; raise onboarding satisfaction from 70% to 85%; cut support tickets per new customer by 30%.
In the real world: Intel’s Andy Grove introduced OKRs, and Google adopted them early; they are now popular in start-ups and tech teams to tie team goals to company priorities. DORA metrics for engineering teams are a similar attempt to measure outcomes.
3. Appraisal methods
| Method | How it works | Caution |
|---|---|---|
| Graphic rating scale | Rate traits or behaviours on a scale | Subjective; vague traits |
| Ranking and paired comparison | Order employees best to worst | Hard in large groups; no absolute standard |
| Forced distribution | Fit ratings to a curve (e.g. 10-20-70) | Can force unfair labels |
| Critical incident | Record specific good and poor events | Needs diary discipline |
| BARS | Rating scale anchored by behaviour examples | Costly to build; very clear |
| Essay and narrative | Written summary | Depends on writing skill |
| MBO / goal-based | Compare results with agreed goals | Needs good goals |
| 360-degree feedback | Input from manager, peers, subordinates, customers and self | Needs trust and anonymity; best for development |
| Assessment centre | Simulations, exercises and tests | Expensive; strong for potential |
| Competency-based | Rate against role competencies | Needs a good competency model |
Competencies combine knowledge, skills, attitudes and behaviours that predict superior performance; competency mapping links them to hiring, training and promotion.
4. Errors and fairness in rating
| Rater error | Description |
|---|---|
| Halo / horn | One strong or weak trait colours everything |
| Leniency / strictness | Rating everyone too high or too low |
| Central tendency | Giving everyone the middle rating |
| Recency | Remembering only the last few weeks |
| Similar-to-me / bias | Favouring those like oneself |
| Contrast | Comparing people with each other instead of the standard |
Remedies: train raters, use behaviour-based criteria, keep continuous notes, use multiple raters, hold calibration meetings, and provide an appeal route. Legal and ethical points: relevant, job-related criteria, consistency, documentation and no discrimination.
5. Feedback, coaching and development
Feedback works when it is timely, specific, behaviour-based, balanced and focused on future action, not on character. A common structure is SBI: Situation, Behaviour, Impact. In the review conversation: prepare, let the employee speak first, discuss evidence, agree on actions, and document. Coaching is ongoing guidance on how to improve; mentoring is a longer-term relationship with a more experienced person.
Handling poor performance: find the cause (skill, will, resources or fit), give clear expectations and support, set a performance improvement plan (PIP) with milestones, and decide on outcome with fairness. High performers need stretch assignments, recognition and career paths.
In the real world: managers who give feedback only once a year leave employees surprised; regular one-to-ones and check-ins reduce surprises and lift engagement.
6. Linking performance to rewards
| Link | Example |
|---|---|
| Merit increase | Raise based on rating |
| Variable pay and bonus | Payout tied to individual, team and company results |
| Incentives and commissions | Sales targets |
| Promotion and career moves | Based on performance and potential |
| Recognition | Awards, public credit |
Cautions: money can crowd out intrinsic motivation; a measure that is rewarded gets optimised, even at the cost of the real goal (Goodhart’s law), so balance quantity with quality and individual with team measures. A balanced scorecard approach uses financial, customer, process and learning measures.
7. Modern redesign of performance management
Annual forced rankings have been widely questioned for harming collaboration and fairness. Several large companies redesigned their systems.
In the real world: GE under Jack Welch was famous for its forced-distribution ranking, then moved away from it in the mid-2010s. Microsoft ended its stack-ranking system in 2013. Adobe dropped annual ratings in 2012 in favour of frequent “check-ins”. Deloitte redesigned its process around frequent conversations, after finding the annual cycle consumed vast time with little value.
Directions today: continuous feedback and frequent check-ins, team-based goals, strengths and growth focus, peer feedback tools, data and analytics to spot bias, and the careful use of AI to summarise performance data, always with human judgement and fairness checks. Remote work needs clear outcome-based measures, since presence is not a signal.
One-minute recap
| Topic | Key idea | Where you meet it |
|---|---|---|
| PM cycle | Plan, monitor, review, reward; more than the annual review | HR systems |
| Goals and measures | KRA, KPI, SMART, MBO, OKRs | Quarterly planning |
| Appraisal methods | Ratings, BARS, 360, assessment centres | Reviews |
| Rater errors | Halo, leniency, central tendency, recency | Calibration |
| Feedback and coaching | SBI model; PIPs; regular one-to-ones | Manager conversations |
| Pay link | Merit, bonus, promotion; Goodhart’s law | Reward design |
| Modern redesign | Continuous feedback; end of forced ranking | GE, Microsoft, Adobe |
Next in MBA Revisited: Human Resource Development.
Written by Claude, Anthropic’s AI assistant, as part of MBA Revisited, a quick refresher of the PGDM subjects one at a time.