Managing civilisation means knowing whether institutions are actually producing the outcomes they were created to deliver. Plans, budgets and projects can all look active while real performance drifts. The professional language includes performance management, monitoring and evaluation, key performance indicators, KPI management, performance measurement, programme evaluation, outcome measurement, performance budgeting, dashboards, benchmarking, evidence-informed decision-making and continuous improvement.
The OECD treats monitoring and evaluation as central to evidence-informed public management because they help organisations understand whether policies and programmes are being implemented as intended and whether they are producing desired outcomes. Monitoring tracks what is happening during implementation. Evaluation investigates whether an intervention worked, why, for whom and at what cost.
The civilisation-level lesson is simple: activity is not the same as performance. Building clinics is activity. Better access to healthcare is an outcome. Training staff is activity. Lower error rates may be an outcome. A management system needs both operational measures and outcome evidence so institutions can learn rather than merely count.
The 60-second answer: what does performance management do?
Performance management converts goals into observable measures, reviews results, investigates variance and adjusts resources or processes when evidence shows that performance is off track. Monitoring provides regular feedback. Evaluation provides deeper analysis of effectiveness and impact. Together they connect strategy, budgets, operations and learning.
- Define outcomes before selecting indicators.
- Measure inputs, processes, outputs and outcomes separately.
- Use leading indicators to detect problems early.
- Use lagging indicators to confirm final results.
- Review trends rather than react to every fluctuation.
- Investigate causes when targets are missed.
- Evaluate programmes when causality and value need deeper study.
- Link performance evidence to resource and management decisions.
- Retire metrics that no longer represent the real objective.
- Preserve lessons so the next cycle starts with better evidence.
Inputs, activities, outputs and outcomes
Performance systems become clearer when they distinguish four levels. Inputs are resources such as money, staff and equipment. Activities are the work performed. Outputs are the immediate products or services delivered. Outcomes are the changes that matter in the world.
Confusing these levels creates false success. A programme can deliver all planned outputs while the intended outcome remains unchanged.
Key performance indicators
A KPI is a measure selected because it represents something important about performance. Good KPIs are connected to decisions and have clear definitions, owners and data sources.
Not every available number deserves KPI status. A long dashboard can hide the few measures that actually need management attention.
Leading indicators
Leading indicators show conditions likely to influence future outcomes. Examples include overdue inspections, teacher vacancies, queue length, near misses, unresolved defects and supplier concentration.
They are valuable because managers can act before the final outcome deteriorates.
Lagging indicators
Lagging indicators show what has already happened: accidents, final exam results, outages, mortality, project completion or annual financial results.
They anchor management in real outcomes but may arrive too late for prevention, which is why strong systems use both leading and lagging measures.
Targets
Targets turn measures into expectations. They can focus attention, but badly designed targets can create gaming or narrow behaviour.
A target should therefore reflect the purpose and be reviewed when context changes. Hitting a target is not meaningful if the underlying outcome has been distorted.
Metric gaming
People adapt to what is measured. If a call centre is rewarded only for short calls, staff may end conversations before problems are resolved. If a hospital is rewarded only for speed, quality may suffer.
Balanced performance systems combine timeliness, quality, safety, access and outcomes so one measure does not dominate behaviour.
Dashboards
Dashboards summarise performance and help managers detect deviation quickly. They should show trends, definitions, freshness and relevant thresholds.
A dashboard without decision rules becomes visual decoration. Every major signal should connect to a possible action or investigation.
Monitoring
Monitoring is regular observation of implementation and performance. It answers questions such as whether milestones are being met, resources are being used, services are reaching users and risks are increasing.
Monitoring is descriptive. It tells managers what is happening so they can adjust during delivery.
Evaluation
Evaluation asks deeper questions about relevance, effectiveness, efficiency, impact and sustainability. It may compare groups, examine counterfactuals, study implementation mechanisms or combine quantitative and qualitative evidence.
Evaluation is especially useful when important resource decisions depend on whether an intervention actually caused the observed result.
Process evaluation
Process evaluation studies how an intervention was implemented. It can explain why a programme that worked in one location failed in another.
Implementation evidence is important because outcome differences may come from delivery quality rather than the underlying idea.
Outcome evaluation
Outcome evaluation examines whether desired changes occurred for intended users. It should use measures connected to the programme’s logic rather than convenient data alone.
Outcomes can take time, so intermediate indicators may be needed before final effects are visible.
Impact evaluation
Impact evaluation seeks stronger evidence about causality: what difference did the intervention make compared with what would likely have happened otherwise?
Methods vary according to feasibility, ethics and context. The principle is to avoid attributing every observed improvement automatically to the programme.
Benchmarking
Benchmarking compares performance across time, sites or comparable organisations. It can reveal variation and identify practices worth studying.
Benchmarking should account for context. A site serving more complex cases may have different performance even with strong management.
Performance reviews
Regular performance reviews bring decision-makers together to interpret evidence, resolve blockers and agree corrective actions. The purpose is not to interrogate teams for bad numbers.
The best reviews create a learning rhythm: understand the signal, test explanations, choose an action and check whether the action worked.
Performance budgeting
Performance budgeting brings information about results into resource allocation. It helps decision-makers ask whether existing spending is producing expected value.
Performance evidence should inform judgement rather than trigger automatic funding formulas because many outcomes depend on external conditions and long-term effects.
Spending reviews
Spending reviews examine existing expenditure and performance together. They can identify programmes that should be redesigned, expanded, reduced or stopped.
This prevents baseline budgets from becoming permanent simply because they existed last year.
Service-level agreements
Service-level agreements define expected performance such as response time, availability or resolution standards. They make interfaces between teams or suppliers more explicit.
Service levels should reflect user need and realistic capability rather than arbitrary numbers.
Operational reviews
Operational reviews focus on near-term performance: queue growth, incidents, staffing, quality, throughput and unresolved risks.
They complement strategic reviews, which look at longer-term outcomes and resource allocation.
Root-cause analysis after missed targets
A missed target is a signal, not an explanation. Managers should investigate whether demand changed, capacity was insufficient, data were wrong, a process failed or the target itself was unrealistic.
Corrective action should address the mechanism rather than simply demand that teams “improve the number.”
Data quality
Performance management is only as reliable as its data. Inconsistent definitions, missing records and delayed updates can make dashboards misleading.
This creates a direct dependency on data governance and shared metadata.
Equity and distribution
Average performance can hide groups who receive much worse outcomes. Managers should examine distribution where relevant: geography, age, service type or other legitimate categories.
A programme can improve the average while leaving severe pockets of failure unresolved.
Qualitative evidence
Not every important outcome is captured by a number. Interviews, observations, case reviews and user feedback can reveal mechanisms and unintended effects.
Strong evaluation combines forms of evidence appropriate to the question instead of assuming quantitative data is always sufficient.
Unintended consequences
Interventions can create effects that were not part of the original objective. Monitoring should therefore include channels for surprises, complaints and anomalies.
Evaluation becomes more useful when it looks beyond the intended KPI set.
Learning loops
Performance systems should shorten the distance between evidence and correction. A measure collected once a year cannot help a process that changes every week.
Review cadence should match how quickly the underlying condition changes and how quickly management can respond.
Worked example: school attendance
A school notices falling attendance. The final exam result is a lagging indicator, while absence patterns provide earlier warning. Staff examine transport, illness, family circumstances and timetable factors before choosing interventions.
Monitoring continues to test whether attendance improves and whether learning follows.
Worked example: maintenance backlog
A utility tracks the number of open work orders. The total backlog appears stable, but critical overdue inspections are increasing.
The performance system is redesigned to separate safety-critical backlog from routine work, producing a measure that better reflects risk.
Worked example: permit-processing time
A city targets faster permit decisions. Average cycle time falls, but rework rises because incomplete applications are being pushed through quickly.
The system adds first-pass quality and applicant resubmission rates to balance the speed target.
Worked example: public-health programme
A vaccination programme tracks doses delivered, coverage by population, missed appointments, stockouts and disease outcomes.
Outputs show whether activity occurred; evaluation investigates whether access and health outcomes improved where intended.
A practical performance-management checklist
- Purpose: What outcome are we trying to improve?
- Logic: How should activities lead to that outcome?
- Measures: Which inputs, outputs and outcomes matter?
- Definitions: Are indicators consistently defined?
- Leading signals: What warns us early?
- Lagging signals: What confirms final performance?
- Targets: Are expectations realistic and useful?
- Distribution: Are some groups or locations performing much worse?
- Review: Who examines results and how often?
- Action: What happens when performance deviates?
- Evaluation: Which questions need deeper causal analysis?
- Budget: Does performance evidence influence resources appropriately?
- Gaming: Are measures creating harmful incentives?
- Learning: Are findings changing policy, process or design?
Common failure patterns
1. Counting activity instead of outcomes
Teams report how much work happened without showing whether the problem improved.
2. Too many indicators
Managers drown in data and miss the few signals that require action.
3. Targets that create gaming
People optimise the measure while degrading the real service.
4. Evaluation done too late
Evidence arrives after key funding or design decisions have already been made.
5. Averages hide inequality
Overall performance improves while specific locations or groups continue to fail.
6. No corrective action
Dashboards identify problems but ownership and response are unclear.
7. Poor data quality
Management reacts confidently to measures that are stale or inconsistently defined.
8. Lessons not reused
Evaluations are published but future programmes repeat the same design mistakes.
How performance management connects to the wider eduKateSG ecosystem
For the broad Civilisation map, use Learn Civilisation with eduKateSG and the Civilisation OS case archive. Performance management sits between strategic planning, operations and public financial management.
It also relies on data governance because indicators are only useful when the underlying data is trusted.
External reference points
- OECD: Public policy monitoring and evaluation
- OECD: Implementation Toolkit for Public Policy Evaluation
Frequently asked questions
What is performance management?
Performance management is the system used to define expected results, measure progress, review evidence and improve performance over time.
What is the difference between monitoring and evaluation?
Monitoring tracks implementation and performance continuously or regularly. Evaluation conducts deeper analysis of effectiveness, impact and why results occurred.
What is a KPI?
A key performance indicator is a measure selected because it provides important evidence about whether a process, programme or organisation is performing as intended.
Why can targets be harmful?
If a target represents only one part of the objective, people may optimise the number while reducing quality or shifting problems elsewhere.
What is performance budgeting?
Performance budgeting uses information about objectives and results to support budget decisions alongside financial and policy considerations.
Conclusion: civilisation must measure what it wants to improve
Without feedback, management becomes assumption extended through time. Monitoring shows whether implementation is on course. Evaluation shows whether the intervention actually worked. Performance management turns both into action.
Managing civilisation therefore means designing evidence systems that are useful enough to change decisions. The goal is not more dashboards. It is shorter learning cycles, clearer accountability and better allocation of scarce effort toward outcomes that matter.