Performance Management

Definition. Performance management (PM) is the continuous process of aligning individual work with organisational goals, setting expectations, observing and supporting performance, evaluating it, and using the evaluation for decisions about pay, development and careers.

Performance management versus performance appraisal. A performance appraisal is a periodic evaluation event, usually annual, that produces a rating or assessment. Performance management is the whole system around it: goal setting, ongoing feedback, coaching, review, and the decisions that follow. Many failures attributed to “performance management” are really failures of an appraisal-only approach, in which nothing happens between one annual form and the next.

Purposes

The purposes conflict. An employee who knows that a conversation will determine their bonus is less likely to discuss weaknesses openly, which undermines the developmental purpose. This tension explains why many organisations separate development conversations (frequent, informal) from evaluation decisions (periodic, calibrated).

First principles: why PM can improve performance

• Goal-setting theory (Locke and Latham): specific and challenging goals lead to higher performance than vague goals such as “do your best”, provided people are committed to the goals, have the ability to achieve them and receive feedback on progress. This is the strongest-evidenced theory in the field and the basis for SMART goals and OKRs.
• Feedback research (Kluger and DeNisi, 1996, a review of over 600 studies): feedback improves performance on average, but in more than a third of cases it reduced performance, particularly when it focused attention on the person (“you are careless”) rather than on the task (“this report missed two data sources”). Well-designed PM therefore emphasises task-focused, specific and timely feedback.
• Expectancy theory (Vroom): people are motivated when they believe effort leads to performance, performance leads to rewards, and they value those rewards. A PM system whose ratings do not visibly affect rewards weakens the second link.

Roles. The employee owns their goals and development and seeks feedback. The manager sets expectations, coaches, evaluates and explains decisions; the quality of PM depends more on managers than on any form or software. HR designs the system, trains managers, runs calibration, checks fairness and connects outcomes to reward and talent processes.

The three pillars named in the compendium remain a useful test of any system: clear and measurable goals and competencies, concrete actions to achieve them, and consequences that follow performance. When consequences are missing, PM becomes paperwork.

Definitions

Key result area (KRA). A broad area of responsibility in which a role is expected to deliver results, such as “workforce stability” or “statutory compliance”. The term is widely used in Indian organisations.

Key performance indicator (KPI). A quantifiable measure of performance in a KRA, with a target. For the KRA “workforce stability”, a KPI could be “monthly voluntary attrition of operators below 1.5%”.

Objectives and key results (OKR). A goal framework consisting of an objective, a qualitative statement of what is to be achieved, and three to five key results, measurable outcomes that show whether the objective has been met. OKRs were developed by Andy Grove at Intel in the 1970s and introduced to Google by John Doerr in 1999.

Comparison

Example for a plant HR manager.

SMART goals

A goal should be Specific (clear about what is to be achieved), Measurable (progress can be tracked), Achievable (stretching but realistic), Relevant (aligned with team and company priorities) and Time-bound (has a deadline). SMART is the practical application of goal-setting theory.

Cascading, alignment and weights

Goals are usually cascaded from company to function, team and individual, so that each person can see how their work contributes. Pure top-down cascading is slow and can produce goals the employee does not own; many organisations now combine it with bottom-up goals proposed by employees, as in MBO. Each goal carries a weight (for example, 30% cost, 30% safety, 20% quality, 20% people development) so that the overall evaluation reflects priorities. Best practice is 4–6 goals per person, not 15.

A competency is an observable combination of knowledge, skills, abilities and behaviours needed for effective performance in a role, defined at different proficiency levels (Chapter 4). In PM, competencies describe the expected behaviours: for example, “Stakeholder management: builds agreement between plant and corporate functions on conflicting priorities”.

Assessing only results rewards people who achieve targets in damaging ways: by cutting safety corners, mistreating teams or hoarding information. Assessing only behaviours rewards people who are pleasant but do not deliver. Many organisations therefore rate the what (results against goals) and the how (competencies and values) separately, and combine them through a matrix.

Why it matters. What leaders reward defines the real culture (Schein, Chapter 7). If a high-revenue salesperson who bullies colleagues receives the top rating, every employee learns that results excuse behaviour.

Overview

Management by objectives (MBO)

Introduced by Peter Drucker in The Practice of Management (1954). The manager and employee agree specific objectives, the employee chooses how to achieve them, progress is reviewed periodically, and evaluation is based on achievement. Its strength is participation and clarity: people commit more to goals they helped set. Example: an SEO manager and marketing head agree to “increase organic traffic by 20% by the end of the third quarter”; the manager chooses the tactics and reports monthly; the year-end evaluation compares actual growth with 20%. Intel’s use of MBO, refined by Andy Grove, evolved into OKRs.

Balanced scorecard (BSC)

Developed by Robert Kaplan and David Norton (Harvard Business Review, 1992) to address the limitation that financial measures describe the past but not the drivers of future performance. It measures performance from four linked perspectives:

The perspectives form a cause-and-effect chain: trained and engaged employees (learning) improve processes (internal), which satisfy customers (customer), which drives financial results. Example from the compendium: an airline trained ground crews to cut aircraft turnaround time, which improved on-time departures, which raised passenger loyalty, which increased revenue. In HR, the BSC is used to set leadership and unit goals and to build HR’s own scorecard.

Behaviourally anchored rating scales (BARS)

BARS replace vague scale points with descriptions of actual behaviour. Building one involves collecting critical incidents from experienced staff, grouping them into performance dimensions, and assigning each incident to a scale point.

BARS improve consistency between raters and give clear feedback, because the employee can see exactly what a higher rating looks like. Their cost limits them to roles with many incumbents.

360-degree feedback

Feedback is collected from several sources: manager, peers, direct reports, the employee themselves, and sometimes internal or external customers. It is most valuable for leadership development, because direct reports and peers see behaviours that managers do not. Best practice: use it for development rather than pay; guarantee anonymity with a minimum number of raters per group (often three); focus questions on observable behaviours; and give the results through a trained coach with a development plan. Its weaknesses are rater fatigue, friendship or rivalry effects, large amounts of data, and confidentiality concerns, particularly in high power-distance cultures where criticising a manager feels risky (Chapter 7).

The 9-box grid

The 9-box grid is a talent-review tool rather than an appraisal method. It places employees on two axes, performance (current results) and potential (likely ability to succeed in larger roles), each at three levels.

It is used in succession planning and to decide where to invest development. Its weaknesses are that potential is hard to judge and prone to bias (managers tend to see potential in people similar to themselves) and that it depends on the quality of performance ratings.

Rating scales

Forced distribution

Under forced distribution (the “bell curve”), a fixed percentage of employees must receive each rating, such as 10% top, 70% middle and 20% bottom. Jack Welch popularised it at General Electric as the vitality curve (top 20%, vital 70%, bottom 10%), with the bottom 10% often exited, a practice known as “rank and yank”.

Microsoft dropped stack ranking in 2013, and GE, Accenture and Infosys moved away from forced curves around 2015–16. Many organisations now use guided distribution: a recommended distribution applied at the level of a large population (a function or business), not imposed on each small team, and adjusted through calibration.

Normalisation is the Indian term for adjusting ratings across teams to fit a distribution or a budget; when done mechanically, it is a frequent cause of employee distrust.

Calibration

Calibration is a meeting in which managers review proposed ratings together so that the same standards apply across teams.

1. Managers submit proposed ratings with evidence (results against goals, examples of behaviour).
2. HR prepares data: proposed distribution by team, gender, location and tenure; prior ratings; outliers.
3. Managers present their top and bottom ratings and borderline cases; peers ask questions and challenge.
4. Ratings are adjusted where standards differ, and the reasons are recorded.
5. HR checks the final distribution for patterns that suggest bias.
6. Managers communicate final ratings to employees and explain the reasoning.

Benefits: consistency, reduced leniency, shared standards, better data for pay decisions. Risks: the most vocal manager’s view dominates; employees not known to other managers are disadvantaged. HR’s role is to facilitate with evidence and to watch for these effects.

Research on ratings has found that a large share of the variation in performance ratings reflects the individual rater’s tendencies rather than the performance of the person rated. A study by Scullen, Mount and Goff (2000), cited in Deloitte’s redesign, attributed about 62% of rating variance to such “idiosyncratic rater effects”.

Controls. Specific goals and behaviourally anchored criteria; evidence gathered through the year (check-in notes, critical incidents); calibration; rater training; and analysis of rating outcomes by gender, location, work arrangement and manager to detect patterns.

Principles of effective feedback, drawn from the research in Section 3.1: it is specific (refers to particular actions and results), timely (close to the event), task-focused (about the work, not the person’s character), balanced (recognises what works as well as what does not) and forward-looking (what to do next).

The SBI model (Center for Creative Leadership) structures feedback in three parts: Situation (when and where), Behaviour (what the person did, observably), Impact (the effect on results or others).

Example: “In Tuesday’s union meeting (situation), you summarised management’s proposal before responding to the demands (behaviour). That lowered the tension and got us an agreement on the shift roster (impact).”

Feedforward (Marshall Goldsmith) focuses on suggestions for the future rather than criticism of the past, which reduces defensiveness.

Check-ins are short, frequent conversations (monthly or quarterly) about priorities, progress, obstacles and development. They do not replace evaluation but make it more accurate and less surprising. Coaching conversations often follow the GROW model: Goal (what the person wants to achieve), Reality (the current situation), Options (possible actions), Will (what they commit to doing).

Manager capability is the constraint. Every redesign that replaced annual reviews with check-ins depended on managers being willing and able to hold good conversations. Organisations that changed the process without training managers found that check-ins simply did not happen.

Diagnosis

Under-performance has three broad causes, and each needs a different response.

IDP and PIP

A fair PIP states the gap with evidence, sets measurable targets that are achievable within the period, provides real support (coaching, training, resources), holds regular reviews with written notes, and states the consequences clearly. A PIP used only as a formality before a decided exit damages trust and the employer brand: Indian employees discuss “PIP culture” openly on professional networks and review sites.

A common error is to treat promotion as a reward for performance in the current role. Performance shows that someone is effective in their present job, but promotion should depend on evidence that they can succeed in the next one. Promoting the best salesperson to sales manager without assessing leadership potential is a classic example.

What happened next. Organisations without visible ratings found that pay decisions became harder to explain, and that managers still effectively rated people in private. Many therefore kept or reintroduced a simple year-end rating, while retaining frequent check-ins. The durable lesson is that one annual event cannot serve all of PM’s purposes at once; effective systems separate frequent development conversations from periodic, calibrated evaluation and from talent reviews.

Technology and AI. PM platforms (SAP SuccessFactors, Workday, Darwinbox, Keka) support goal tracking, check-ins and continuous feedback. AI tools are being used to summarise feedback, draft reviews and flag rating patterns. They can improve evidence and detect bias, but using AI to decide ratings raises fairness, transparency and data-protection concerns (Chapter 9).