Monitoring that leads to decisions: designing MEL systems people actually use

Most programmes have a monitoring, evaluation and learning (MEL) plan. Far fewer use it to change anything. Reports arrive late, indicators number in the hundreds and the data ends up in an annex. The fix is not more data. It is designing the system around the decisions it needs to support.

Start with the decisions

Before choosing indicators, ask who needs to decide what, and when. A district manager choosing where to send supervisors, a department deciding whether to scale a pilot, and a donor deciding whether to extend funding all need different information at different times.

Keep the framework small

A clear theory of change shows how activities are expected to lead to results. From it, select a small set of indicators that track the critical steps. For each one, define a baseline, a target, a data source, how often it is collected and who owns it.

A useful test for every indicator:

• Is it needed for a decision someone will actually make?
• Can it be measured reliably at a reasonable cost?
• Will someone act on the result?

If the answer to any of these is no, drop it.

Use routine data where you can

Health information systems already collect large volumes of data. Strengthening them is usually better value than parallel data collection, and it leaves a lasting system behind. That means investing in data quality assessments, clear definitions and feedback to the people who enter the data.

Use surveys, facility assessments and qualitative methods such as key informant interviews and focus groups for what routine data cannot show.

Build learning loops into the calendar

Learning does not happen by accident. Schedule short review meetings where teams look at results and agree what to change. Plan a mid-term evaluation while there is still time to adapt, and an end-line evaluation that uses mixed methods to explain not only what changed but why.

Communicate for the audience

A ten-page policy brief often does more than a hundred-page report. Present findings in plain language, with a clear recommendation and a named owner for each action.

MEL works when it feels like part of management rather than an external obligation. That is the standard we aim for when designing frameworks and assessments for clients.