Monitoring

“Monitoring tells us what is happening often enough to respond; it does not tell us automatically why it happened. ”

Monitoring is the routine evidence system of sustainability practice. It tracks whether activities occur, controls function, conditions change and targets remain on course. Because it operates continuously or periodically, it can identify problems earlier than an endline study or annual audit.

The OECD evaluation glossary defines monitoring as a continuing function that uses systematic data collection on specified indicators to provide management and stakeholders with information on progress, achievement of objectives and use of resources. The emphasis is not collection alone. Monitoring includes analysis, communication and response. Satellite alerts illustrate the distinction.

A forest-monitoring system can detect a possible canopy disturbance quickly. The alert is valuable because it directs attention.

It may not reveal whether the change is deforestation, harvesting, fire, cloud artefact or lawful land use. Verification and contextual data are still required. Good monitoring begins with a decision. What would the organisation do if the value moved? If there is no threshold, owner or response, the indicator can become reporting burden rather than management evidence.

Dashboards filled with data that no one acts upon are archives, not monitoring systems. Frequency should reflect the speed and consequence of change. Worker safety incidents may require immediate notification. Groundwater trends may need seasonal interpretation. Farmer income may be measured annually but supplemented with price and cost monitoring during the season.

More frequent data are not automatically better when noise exceeds meaningful change. Monitoring should include implementation and outcomes.

Tracking training delivered can reveal schedule and coverage. Tracking adoption, persistence and effects tests whether delivery is leading towards intended change. The results chain helps prevent outputs from standing in for outcomes. Qualitative monitoring matters where experience, trust or power cannot be inferred from counts.

Regular worker interviews, community observation and grievance analysis can reveal conditions hidden by formal records. Repetition should not expose participants to retaliation or survey fatigue. Risk-based monitoring allocates attention where harm is more likely or severe. It can increase efficiency, but it should not create blind spots.

Low-data areas may be classified as low risk simply because the system has little evidence. Some baseline coverage remains necessary to detect change and model error.

Automated monitoring can scale response but introduces new governance needs. Thresholds create false positives and negatives. Models drift. Sensors fail. People affected by alerts may need a route to correct the record. Human review should be proportionate to consequence. Monitoring is not evaluation. A trend can show that an outcome changed but not establish why.

External prices, weather, policy or other programmes may contribute. Monitoring supports causal questions; it does not resolve them by itself. The process should adapt. If the same data are collected for years without informing action, the system should change. If new harms emerge, indicators and channels should be added. Monitoring is part of continuous improvement, not a permanent questionnaire.

Monitoring architecture should match the speed and severity of the issue.

Satellite alerts may be reviewed weekly, serious safety incidents immediately and household-income surveys seasonally. Collecting every indicator at the same frequency wastes resources and can still miss the moment when intervention matters. Escalation rules should distinguish routine variation, warning signals and conditions requiring urgent action. The burden of monitoring is part of its design.

Farmers, workers and local organisations are often asked for the same information by multiple buyers and schemes without receiving useful feedback. Data collection that consumes time, creates risk or duplicates existing records can undermine the relationship it is meant to improve. A responsible system minimises burden, returns findings to contributors and collects information because a decision will be made with it.

The discipline is to connect every signal to interpretation and response.

What constitutes normal variation? What triggers investigation? Who decides? How quickly? Monitoring becomes useful when evidence changes behaviour before the annual report describes what already happened.

Practical application

Define indicators, frequency, thresholds, owners and response protocols. Combine administrative, remote, qualitative and direct evidence, and monitor data quality alongside performance. Include early-warning and outcome measures.

Review false alerts, missed events and burden. Give affected people routes to challenge records. Periodically retire indicators that do not inform decisions and add measures for emerging risks. Build a signal-to-action matrix before collection begins. For every indicator, specify normal range, warning threshold, critical threshold, responsible reviewer, investigation step and response time.

Periodically test whether alerts led to action and whether action improved the underlying condition. Monitoring quality should be judged partly by decisions changed, not only by records collected. Archive alert decisions as well as data so later reviewers can see when evidence was available, how it was interpreted and whether delay contributed to harm or missed opportunity.

Why it matters

Sustainability conditions can change faster than audit or reporting cycles. Monitoring allows earlier prevention, adaptation and corrective action, but only when signals are connected to accountable response.

Common misconception

Monitoring is often treated as routine data collection or proof of impact. It tracks what is happening against expectations; causal explanation and higher-level judgement require evaluation and other evidence.

Connections

Indicator and Metric define what is tracked. Target and Baseline provide reference points. Evaluation interprets relevance, performance and causality, while Continuous Improvement changes the system in response to what monitoring reveals.

A question worth asking

Which monitored signal would trigger a real operational decision tomorrow, and which ones exist only because the report expects them?

Selected references

OECD. 2023. Glossary of Key Terms in Evaluation and Results-Based Management for Sustainable Development, Second Edition.

United Nations Development Programme. 2009. Handbook on Planning, Monitoring and Evaluating for Development Results. Kusek, J. Z. and Rist, R. C. 2004. Ten Steps to a Results-Based Monitoring and Evaluation System. Food and Agriculture Organization of the United Nations. 2017. Voluntary Guidelines on National Forest Monitoring. Patton, M. Q. 2011. Developmental Evaluation.