Monitoring & Evaluation for NGOs: Indicators, Baselines and Learning Loops

Monitoring and evaluation is often treated as a donor requirement: a table to fill in at proposal stage and a report to produce at the end. Used well, it is something far more valuable. Good monitoring and evaluation tells your team, while there is still time to act, whether your work is making the difference you intended, and why or why not. This guide walks through the building blocks: indicators, baselines and targets, means of verification, the OECD DAC evaluation criteria, a simple M&E plan, and the learning loops that turn data into better decisions.

Monitoring and evaluation: what is the difference?

The two terms are usually spoken together, but they do different jobs.

	Monitoring	Evaluation
Purpose	Track progress and flag problems during implementation	Judge the value, results and lessons of the project
Timing	Continuous: weekly, monthly, quarterly	Periodic: mid-term, final, sometimes ex-post
Main questions	Are we doing what we planned? Are we on track?	Are we doing the right things? What changed, and why?
Who does it	Project staff, partners, sometimes participants	Internal team, external evaluator or both
Focus	Activities, outputs, early outcomes	Outcomes, impact, relevance, sustainability

Both depend on the same foundation: a clear statement of what the project is trying to change. If your objectives are vague, no amount of data collection will rescue your M&E. That is why monitoring and evaluation design starts with the results chain or Logical Framework, which our Project Management Fundamentals workshop covers as part of defining a project and its objectives.

Start with the results chain

Most NGO projects can be described as a chain: inputs (money, staff, materials) enable activities (training, outreach, service delivery), which produce outputs (people trained, materials distributed), which lead to outcomes (changes in knowledge, behaviour, practice or conditions), which contribute to impact (long-term change in people’s lives or systems). Each link rests on assumptions that may or may not hold.

Indicators sit at each level of this chain. The common trap is to measure only what is easy, which usually means activities and outputs. Counting workshops tells you that you were busy; it does not tell you whether anything changed.

Writing SMART indicators

An indicator is a specific, observable measure that shows whether change is happening. The SMART test is a useful filter:

• Specific: it describes exactly what is measured and for whom.
• Measurable: it can be counted, rated or otherwise assessed consistently.
• Achievable: the target is realistic given resources and time.
• Relevant: it genuinely reflects the result it sits under.
• Time-bound: it states when the target should be reached.

From weak to strong: a worked example

Consider a hypothetical project supporting smallholder farmers to adopt water-saving techniques.

Weak indicator	What’s wrong	Stronger version
Farmers’ awareness improved	Not measurable; no group or timeframe	Percentage of participating farmers who can correctly describe at least three water-saving techniques, by end of month 6
Number of trainings held	Activity, not a result	Percentage of trained farmers using at least one technique on their plot one season after training
Improved livelihoods	Too broad; many causes	Change in average yield per hectare among adopting farmers compared with baseline, by end of year 2

Disaggregate from the start

Decide at design stage which groups you need to see separately: for example by gender, age, disability or location. Disaggregation is almost impossible to add later if the data was never collected. It is also how you notice when a project works well for some people but leaves others behind.

Baselines and targets

A baseline is the value of an indicator before your intervention begins. A target is the value you aim to reach by a given point. Without a baseline, a target is a guess and a result is impossible to interpret. If 60% of farmers use a technique at the end of the project, that is a success if the baseline was 10% and a non-event if it was 55%.

1. Collect baselines early, ideally before activities start, using the same method you will use later.
2. Use existing data where it is reliable, such as government statistics or your previous projects, and note its source and date.
3. Set targets with the people who will deliver them, based on the baseline, past experience and resources.
4. Add milestones (interim targets) so you can tell during the project whether you are on track.
5. Record changes to targets and the reason, and agree them with the donor where required.

Means of verification

Means of verification are the sources of evidence that show an indicator’s value: where the data comes from and how it is gathered. In a logframe they appear in their own column. Common sources include:

• Attendance sheets, registration databases and distribution lists (for outputs).
• Pre- and post-tests, skills assessments and observation checklists (for learning).
• Household or participant surveys at baseline, mid-term and endline (for outcomes).
• Focus group discussions and key informant interviews (for explaining why change did or did not happen).
• Official records, such as school enrolment or health facility data (for wider outcomes).
• Most significant change stories and participatory tools (for unexpected and qualitative change).

Balance rigour with cost. A large household survey may be the gold standard, but if it consumes most of your M&E budget, a smaller sample combined with good qualitative work may tell you more. Always respect data protection: collect only what you need, explain to participants why you collect it, store it securely and follow GDPR where it applies.

Bringing it together: a simple M&E plan

An M&E plan puts all of this into one place. The format below works for most small and medium projects and can be expanded for larger programmes.

Result level	Indicator	Baseline	Target	Means of verification	Frequency	Responsible	Use of data
Outcome	% of trained farmers using at least one water-saving technique	Baseline survey, month 1	Set after baseline	Follow-up survey and field observation	Each season	M&E officer	Quarterly review; adjust follow-up visits
Output	Number of farmers completing the full training (by gender)	0	As per proposal	Attendance database	Monthly	Field coordinators	Monthly team meeting; donor report
Output	% of participants passing post-training assessment	Pre-test score	Agreed with trainers	Pre/post tests	Each cohort	Lead trainer	Revise training content

The final column, use of data, is the one most plans leave out, and it is the most important. If you cannot say who will use an indicator and for what decision, consider dropping it.

Evaluation and the OECD DAC criteria

The OECD Development Assistance Committee’s evaluation criteria are the most widely used framework for evaluating development and humanitarian work. Many donors expect evaluations to be structured around them. Since their revision in 2019 there are six:

Criterion	Core question	Example evaluation question
Relevance	Is the intervention doing the right things?	Did the project respond to the priorities of the farmers it served?
Coherence	How well does it fit with other interventions?	Did it complement or duplicate government extension services?
Effectiveness	Is it achieving its objectives?	To what extent did adoption of techniques increase, and for whom?
Efficiency	How well are resources being used?	Were results delivered on time and at reasonable cost?
Impact	What difference does it make?	What longer-term changes, intended or not, can be observed?
Sustainability	Will the benefits last?	Are farmers still using the techniques after support ended?

You do not need to give every criterion equal weight. A good evaluation’s terms of reference prioritise the questions that matter most for the decisions ahead, such as whether to scale up, adapt or close a programme. Humanitarian organisations may also assess their work against the commitments of the Core Humanitarian Standard.

Learning loops and adaptive management

Data only creates value when someone uses it to make a decision. A learning loop is a deliberate, regular cycle that moves from data to reflection to action.

1. Collect: gather monitoring data as planned.
2. Make sense: analyse it with the team, looking for patterns, surprises and gaps.
3. Reflect: ask “so what?” Are our assumptions holding? Who is being missed?
4. Decide: agree specific adjustments, owners and deadlines.
5. Act and record: implement the changes, note them in a decision log and track their effect.

In practice this can be a quarterly two-hour “pause and reflect” session with programme, finance and M&E staff, and, where possible, partners and participant representatives. Use three questions: What is working? What is not? What will we change?

Adaptive management means being willing to change activities, and sometimes outputs, in response to what you learn, while staying accountable for outcomes. Most donors accept well-argued adjustments, especially when they are backed by monitoring evidence and agreed through a formal change request. Managing those changes cleanly, along with schedule and budget tracking, is a core topic in our Advanced Project Management Tools & Techniques workshop.

A hypothetical example

Imagine the farmer project’s first-season monitoring shows that men are adopting techniques at a much higher rate than women. A reflection session with field staff reveals that training is held at times when many women are collecting water, and that demonstration plots are on land that women rarely visit. The team moves sessions to different hours, adds demonstrations near community water points and records the decision. The next season’s data shows the gap narrowing. Without disaggregated indicators and a regular learning session, the problem would likely have surfaced only in the final evaluation.

Common mistakes to avoid

• Too many indicators. Data that nobody uses wastes participants’ time and staff effort.
• Measuring only outputs. Activity counts cannot show change.
• No baseline. Results become impossible to interpret.
• M&E as one person’s job. Programme staff must own and use the data.
• No budget for M&E. Surveys, analysis and evaluation all cost money; plan for them from the start.
• Ignoring negative findings. Problems found early are cheaper to fix and build donor trust when reported honestly.
• Collecting personal data carelessly. Protect participants’ privacy and dignity at every stage.

Strong monitoring and evaluation grows out of strong project design. Explore our project management courses for NGO staff, held as practical one-day workshops in Brussels, or ask about 1:1 and small-group coaching to help your team build an M&E plan for a live project.