A programme can meet its targets and still fail the people it was designed to serve.
It can report that thousands of households received support, activities were delivered on schedule, participation remained high and expenditure followed plan. Its dashboard may be green. Its quarterly report may be approved. Its indicators may be technically sound.
And yet, somewhere behind those numbers, a mother may have stopped attending because meetings were held at the wrong time. A young person may have remained silent because speaking openly in front of local leaders felt unsafe. A family may have received assistance but experienced the registration process as humiliating. A person with a disability may never have appeared in the monitoring system at all.
None of this necessarily appears in the indicators. Not because it does not matter, but because the system may never have been designed to capture it.
This is one of the central challenges of Monitoring, Evaluation, Accountability and Learning (MEAL): organisations can become very good at measuring what is visible while remaining less capable of understanding what is missing, difficult to express or simply outside the questions they chose to ask.
The illusion of complete data
Data gives organisations something valuable: a way to observe patterns, compare locations, monitor progress and support decisions. But no dataset is a complete representation of reality.
Every monitoring system is shaped by choices: what is measured, who is asked, when information is collected, which categories are created and which experiences are left outside the system. A dataset can therefore be technically accurate and still provide only a partial picture.
Consider a programme that monitors attendance at community sessions. The records show that 80% of expected participants attended. On paper, this looks encouraging.
But who did not attend? Were they working? Did they live too far away? Were they informed too late? Did the meeting take place at a time that conflicted with household responsibilities? Did they feel welcome? Were women able to express disagreement freely? Were people with disabilities able to access the venue?
Attendance data alone cannot answer these questions. The danger begins when organisations confuse what they can measure with everything that matters.
There is also a more subtle problem. Sometimes we know that information is missing. Sometimes we know that certain people are missing from the dataset. But the most difficult situation is when we do not even know that something, or someone, is absent.
What is not designed to be collected can easily disappear from organisational attention. That is not merely a technical issue. It can also become a question of power.
If people cannot influence what is measured, how feedback is collected or how findings are interpreted, they risk becoming visible only as beneficiaries, respondents or numbers in a report, not as people with knowledge, preferences and rights.
Participation is not the same as voice
Development and humanitarian programmes frequently speak about participation. Communities are invited to meetings. People are consulted during assessments. Participants are asked to validate plans.
These practices can be meaningful. But participation should not be confused with influence.
Being present in a meeting does not guarantee that someone feels able to disagree. Answering a survey does not guarantee that the answer will influence a decision. Signing an attendance sheet tells us almost nothing about whether an intervention was relevant, respectful or safe.
The distinction matters because power is rarely distributed equally in a programme.
A young person may hesitate to contradict an adult leader. A woman may attend a meeting without feeling free to speak. A person with a disability may technically be invited but practically excluded. A beneficiary may have a serious concern but remain silent because complaining could, in their perception, jeopardise future assistance.
A programme can therefore be highly participatory on paper while remaining only weakly responsive in practice.
The Core Humanitarian Standard places this issue at the centre of accountable humanitarian action. It states that people affected by crisis should have access to information, participate in decisions that affect them and have safe ways to raise concerns and complaints, without fear of retaliation. Organisations are expected to respond to those concerns and use them to improve their work.
This leads to an important distinction: a feedback mechanism is not accountability by itself.
A suggestion box is not accountability if nobody reads what is placed inside it. A hotline is not accountability if people do not trust it. A community consultation is not meaningful participation if the important decisions have already been made elsewhere.
Accountability begins when people can influence action : and when organisations are prepared to respond.
The people missing from the dashboard
The people represented in monitoring systems are often those who were easiest to reach. They may be regular participants, registered beneficiaries, people with working phones, residents of accessible communities or individuals comfortable speaking with project staff.
But some of the most important information may come from those who never appear in the dataset.
The household that stopped attending after receiving the first service. The adolescent who did not feel safe answering a survey. The person living with a disability who could not access the meeting venue. The farmer whose field is too remote for regular monitoring visits. The family that chose not to complain because it feared losing support.
These absences can reveal something important about programme performance. They may indicate barriers to access, weaknesses in communication, social exclusion or unintended consequences that conventional indicators do not capture.
This matters in Madagascar. Digital access remains highly unequal. DataReportal estimated that 6.71 million people in Madagascar were using the internet at the end of 2025, equivalent to 20.4% of the population. In other words, most of the population was still not using the internet.
The implication is not that digital tools should be abandoned. Quite the opposite. Digital tools can make data collection faster, improve data management and create new possibilities for feedback and analysis.
But digital cannot become the only door through which people are allowed to be heard.
Depending on the context, effective feedback may require a combination of face-to-face conversations, community meetings, telephone channels, focal points, paper-based options or confidential referral mechanisms.
The right question is not “which channel is most modern?” It is: which channels allow the people concerned to communicate safely and meaningfully?
From collecting feedback to closing the loop
Having a feedback mechanism does not necessarily mean having a feedback culture.
A mechanism receives information. A feedback culture ensures that information reaches the people who can act on it, is analysed in time, influences decisions where appropriate and generates a response to those who provided it.
This is the logic behind closing the loop. ALNAP’s guidance on effective feedback mechanisms emphasises the importance of designing feedback systems around the information that programme teams actually need for decision-making, ensuring that feedback reaches decision-makers and using it as part of organisational learning and improvement.
Consider a livelihood programme receiving repeated feedback that training sessions conflict with market days.
The organisation could simply record the complaint in a spreadsheet. It could move the sessions to another day. Or it could investigate further and discover that the problem also involves childcare responsibilities, travel costs or household decision-making.
The first option produces a record. The second may improve attendance. The third creates learning.
This is where Accountability and Learning meet : the same logic behind the ex-USAID CLA framework (Collaborating, Learning and Adapting), where feedback only has value if it actually changes a practice.
Closing the loop does not necessarily mean accepting every request. Some requests may be impossible, unsafe or outside the programme’s mandate. It means responding transparently: what did we hear? What did we do? What could we not change, and why?
That response is part of accountability.
Building systems that can listen
A stronger MEAL system does not require an endless list of questions or a sophisticated digital platform. It requires intentional design.
Before adding another indicator, an organisation should ask: whose experience could we be missing?
Before creating another survey, it should ask: who is unlikely to answer this survey?
Before choosing a digital feedback channel, it should ask: who cannot realistically use it?
In practice, this may mean:
- providing more than one feedback channel;
- creating safe options for sensitive concerns;
- considering the needs of women, young people, people with disabilities and other groups who may face barriers to participation;
- tracking not only the number of complaints received, but also response times, recurring themes and actions taken;
- returning findings to communities in clear and accessible language;
- involving field teams in data interpretation, because they often understand the context behind a trend before it becomes visible in a dashboard;
- combining quantitative monitoring with qualitative inquiry.
None of this means making MEAL unnecessarily complicated. The objective is the opposite: to make the system more honest about what it knows and what it does not know.
From tools to organisational practice
This is where capacity building must go beyond software training.
Excel, KoboToolbox, Power BI, DHIS2 and other platforms can strengthen data collection, storage, analysis and visualisation. They are valuable. But no software can decide which voices deserve attention.
A dashboard cannot determine whether a person felt safe enough to complain. A database cannot decide whether an organisation is willing to change course after receiving uncomfortable feedback. Those are organisational choices.
This is where the intersection between MEAL, data analytics and organisational practice becomes important.
The role of an organisation such as Helpazone should therefore not be limited to teaching professionals how to use individual tools. The more important task is to help teams connect management questions with data needs, collection systems, data quality, analysis, visualisation and decision-making.
This is the logic behind Helpazone’s approach to MEAL, data analytics, information systems and programme management: developing practical capabilities that organisations can use beyond the training room.
The objective is ultimately simple: teams should be able to collect evidence responsibly, understand what it reveals, recognise what remains unseen and act before a problem becomes irreversible.
The measure of a listening organisation
The strongest MEAL systems do not only ask whether a project delivered its activities. They ask whether people were heard. They ask whether those who were absent were noticed. They ask whether feedback changed anything. And they ask whether the organisation was willing to learn from information it did not want to hear.
That is the difference between monitoring people and being accountable to them.
In a sector that increasingly values data, the real test is not how much information an organisation can collect. It is whether it knows the limits of that information. Whether it can recognise who is missing. Whether it can create safe spaces for disagreement. Whether it can turn feedback into learning.
And, ultimately, whether it has built the humility, discipline and courage to listen.
Because sometimes the most important data in a programme is not the number that appears on the dashboard. It is the voice that never made it there.