Designing a MEL plan
that survives the field

I have watched beautiful MEL plans die in a drawer. Impeccable logical frameworks, dozens of indicators, matrices validated by everyone — and six months later, teams filling in late Excel files, incomplete data and donor reports written in a rush. The problem is almost never the theory: it is that the plan was not designed for the real field — the one with patchy connectivity, rotating staff and days that are already too short. Here are the five principles I apply, project after project, so that a monitoring & evaluation system lives beyond its binder.

1. Start from decisions, not indicators

The first question of a MEL plan is not "what should we measure?" but "which decisions should this data inform, and for whom?" A project coordinator, a partner ministry and a donor such as BMZ or DGIS do not need the same information, nor at the same frequency.

Starting from decisions, you quickly discover that a handful of well-chosen indicators, aligned with the logical framework, beats forty "just in case" indicators. Every indicator has a cost: collection, cleaning, analysis. An indicator nobody uses to decide is a tax on your teams.

2. Design for field reality

A MEL plan is designed in an office but executed in areas where connectivity is intermittent, where staff change positions, and where the enumerator sometimes works in a different language from the questionnaire. Ignoring these realities means programming the collection to fail.

• Offline-first collection: tools must work without network and synchronize later.
• Forms in the enumerator's language, tested in the field before deployment, not after.
• Procedures that survive departures: short documentation, recorded trainings, buddy systems — staff turnover is a certainty, not a risk.

3. Digitalize collection without complicating it

Digitalization is not an end in itself: it is a means to get more reliable data, faster. Tools like KoboToolbox or ODK validate data at the source — input constraints, skip logic, dropdown lists — and eliminate weeks of downstream cleaning.

But the golden rule remains simplicity: a well-built 20-minute form will be filled in correctly; a one-hour form will be rushed, whatever the tool. Digitalize the process as it should be, not as it has accumulated over the years — which is precisely why process mapping always precedes digitalization in my practice.

4. Make the data visible, fast

Nothing kills data collection more surely than silence. If field agents send data for months without ever seeing the result, quality drops — why polish what disappears into a black hole?

This is where dashboards change the game. Even a simple Power BI dashboard, automatically refreshed and shared with the teams, turns collection into a feedback loop: everyone sees their contribution, anomalies are spotted in days rather than quarters, and project reviews rely on fresh figures rather than memories.

5. Plan for learning, not just reporting

The "L" of MEL is the first one we sacrifice — yet it is the one that justifies everything else. A system that only produces donor reports is a system half dead.

Build into the plan, from day one: periodic reviews where data is discussed (not just presented), capitalization moments where lessons become reusable documents, and a budget to adjust the tools along the way. A living MEL plan is a plan that is allowed to change.

A MEL plan that survives the field is not the most sophisticated one: it is the one teams take ownership of because it genuinely helps them work. Sophistication can come later — trust in the data must be there from day one.

Merveille Aganze Sami

MEL & Database Management Advisor. 9+ years of experience in monitoring & evaluation, GIS and digitalization with international organizations (GIZ, Enabel) in DR Congo.

LinkedIn Contact me