Sopact is a technology based social enterprise committed to helping organizations measure impact by directly involving their stakeholders.
Copyright 2015-2026 © sopact. All rights reserved.
Monitoring, evaluation and learning (MEL) adds the function that turns evidence into a decision in time. Learn the MEL cycle, framework, and learning agenda.
Monitoring, evaluation and learning (MEL) extends M&E by adding a deliberate learning loop: the practice of feeding monitoring and evaluation evidence back into program decisions while the program is still running, not just reporting it at the end. The first two letters describe what most systems already do; the third is the one that changes the program in time to matter.
That third letter is where Sopact concentrates. Monitoring and evaluation produce evidence; learning is what happens when that evidence comes back early enough to revise the program. Sopact makes learning a property of the system rather than a hope: assumptions monitored mid-cycle, a quarterly review on the calendar, and every revision documented as the record of what the program learned. This page is about the learning loop specifically; for the measurement layer beneath it, see monitoring and evaluation.
Used by: MEL leads, program directors, foundation program officers, and adaptive-management teams who have monitoring dashboards and evaluation reports but no mechanism that turns either one into a decision before the cohort has already graduated.
Ask any funder proposal and it promises learning. Ask the same organization a year later what it changed on the evidence, and the answer is usually nothing, because learning requires the evidence to come back in time to change something, and an annual evaluation cannot do that. By the time the report lands, the cohort it describes is gone. The Learning Gap is this structural mismatch: the letter every organization claims and almost none operationalizes, because the cadence of traditional evaluation is slower than the cadence of the program.
Closing the gap is not a matter of wanting to learn more. It is a matter of timing. Learning becomes real only when three things are true: the program's assumptions are monitored mid-cycle instead of tested once at the end; a review is scheduled often enough to act on what the monitoring shows; and every resulting revision is written down, so the change is a documented decision rather than a hallway conversation. Sopact builds all three into the system: the assumptions come from your theory of change, the review runs quarterly, and the revision log becomes the record of what the program learned.
This is what separates MEL from M&E in practice, not just on paper. M&E can be a reporting function that runs parallel to the program and never touches it. MEL, done properly, is a control loop: monitoring feeds a review, the review changes the program, and the next round of monitoring tests whether the change worked. The evidence and the decisions live in the same loop instead of two separate documents that never meet.
MEL has three parts and one rhythm. Monitoring is the continuous tracking of activities, outputs, and leading indicators while the program runs. Evaluation is the periodic, structured judgment of whether outcomes occurred and whether the program caused them. Learning is the deliberate act of feeding both back into decisions, and unlike the first two it only exists if there is a cadence that forces it to happen.
That cadence is the operational core of MEL. The practice Sopact runs is a quarterly assumption review: each quarter the team takes the monitoring evidence, checks it against the assumptions in the theory of change, decides which assumptions are holding, weakening, or broken, and documents every revision. The assumptions are not invented for the review; they are the same conditions named under each arrow when the theory of change was built, which is why a MEL system built on a theory of change in monitoring and evaluation can point to the exact link that failed instead of just a low number.
This is adaptive management with a paper trail. Adaptive management is the discipline of adjusting a program on evidence as it runs rather than waiting for a final verdict; the quarterly review is where it happens, and the documented revision is what makes it accountable. A MEL framework that skips the documentation loses the learning the moment the meeting ends: the whole point of the revision log is that next year you can see what you changed and why. For the indicator hierarchy the monitoring side reports into, see the results framework guide, and the tooling that runs it in the monitoring and evaluation tools guide.
Watch — the monitoring, evaluation and learning training series. A walk-through of building MEL that runs the learning loop off your theory of change, from mid-cycle assumption monitoring to the quarterly review and the documented revision. Presented by Unmesh Sheth.
MEL and M&E are not competing frameworks: MEL is M&E with an explicit learning loop bolted on, and the difference is entirely in the third letter. M&E answers two questions: is the program being delivered as planned (monitoring), and did it produce the intended outcomes (evaluation). MEL adds a third: given what monitoring and evaluation are showing right now, what should the program change, and it forces that question on a cadence rather than leaving it to the final report. In an M&E system the evidence flows one way, into a report. In a MEL system the evidence flows back into the program while there is still a program to change.
The connection that makes the loop work is to the theory of change. Monitoring tracks indicators, but indicators alone tell you a number moved, not why. When the monitoring data is tied to the assumptions the theory of change names (participants can attend evening sessions, employers value this credential), a falling indicator points to a specific broken assumption, and the quarterly review has something precise to act on. That is how MEL turns monitoring into learning: the assumptions are the bridge between a number that dropped and a decision about what to do. A theory of change with tested assumptions and a quarterly review is, in effect, a working MEL framework; the broader practice sits inside impact measurement & management.
A MEL system earns its keep at four points: collecting on one persistent ID, monitoring the assumptions mid-program, running the quarterly learning review, and adapting the program with the revision documented. The animation below runs the loop; the four prompts under it are the ones behind each job.
1 · Build the backbone the loop runs on. Draft the theory of change with an assumption named under every arrow; those assumptions are what the quarterly review will monitor. The walkthrough is in how to build a theory of change.
Academy walkthrough → How to build a theory of change
Build a theory of change from this program description: [PROGRAM URL OR DOC]. Draft the outcome stages, and under every causal arrow name the assumption the link depends on. Return the assumptions as a monitoring list, each one phrased as a question a MEL review could check each quarter.
2 · Turn the assumptions into monitoring questions. Convert each assumption into a mid-program question with a threshold that triggers a review, so a weakening link surfaces in weeks.
Academy walkthrough → How to audit a theory of change
From this theory of change: [PASTE OR LINK], build an assumption monitoring calendar for a MEL system. For each assumption, write the monitoring question, the instrument and week it belongs in, and the response pattern that signals the assumption is breaking. Rank assumptions by how much the whole theory depends on them.
3 · Wire the results framework so monitoring rolls up. Structure the indicator hierarchy so the quarterly review reads a coherent picture, not scattered metrics.
Academy walkthrough → How to build a results framework
From this theory of change: [PASTE OR LINK], produce a results framework for a MEL system: the results hierarchy with an indicator on each level, the collection point and instrument for each, and confirmation that every indicator links to one persistent stakeholder ID so pre-post is individual, not population-level.
4 · Standardize the fields so the loop is comparable over time. Lock the definitions once with a data dictionary so this quarter's review compares to last quarter's, not to a moved goalpost.
Academy walkthrough → How to build a data dictionary
From this results framework and set of instruments: [PASTE OR LINK], build the data dictionary for a MEL system: for every field, the definition, type, allowed values, instrument, and collection point. Flag any indicator whose definition could drift between quarters and would break longitudinal comparison in the learning review.
The sections above are the argument; the Academy articles are the practice, each a hands-on companion written to run on your own data.
Monitoring, evaluation and learning (MEL) is M&E extended with a deliberate learning loop: monitoring tracks the program continuously, evaluation judges whether outcomes occurred, and learning feeds that evidence back into decisions while the program is still running. The learning component is the one most organizations claim but rarely operationalize. In Sopact, learning is built into the system as a quarterly assumption review with every revision documented, so the evidence changes the program in time to matter rather than arriving in a final report.
M&E answers two questions: is the program being delivered as planned, and did it produce the intended outcomes. MEL adds a third: given what monitoring and evaluation show right now, what should the program change, asked on a cadence rather than at the end. In an M&E system the evidence flows one way into a report; in a MEL system it flows back into the program while there is still a program to change. Sopact operationalizes the difference through a scheduled quarterly review tied to the theory-of-change assumptions.
Learning in MEL means feeding monitoring and evaluation evidence back into program decisions while the program is running, and documenting each resulting revision. It is not a vague commitment to reflection: it requires the evidence to arrive in time to change something, which an annual evaluation cannot do. Sopact makes learning concrete: assumptions are monitored mid-cycle, a review runs every quarter, and every revision is logged as the record of what the program learned.
A MEL framework is the structure that connects monitoring, evaluation, and learning into one loop: the indicators tracked, the assumptions monitored, the cadence of review, and the way evidence feeds back into decisions. A theory of change with tested assumptions, a results framework of indicators, and a scheduled quarterly review is, in practice, a working MEL framework. Sopact builds the framework so the learning loop is enforced by cadence and every revision is documented, not left to intention.
Adaptive management is the discipline of adjusting a program on evidence as it runs rather than waiting for a final verdict. In a MEL system it happens at the quarterly learning review, where the team checks monitoring evidence against the theory-of-change assumptions, decides which are holding or breaking, and revises the program. Sopact makes adaptive management accountable by documenting every revision, so the adjustment is a dated decision on the record rather than an untracked hallway change.
A MEL plan is the written specification of how a program will monitor, evaluate, and learn: which indicators are tracked, which assumptions are monitored and when, the review cadence, and how evidence feeds back into decisions. A strong MEL plan is derived from the theory of change so every indicator and assumption traces to a causal claim. In Sopact the plan is not a static document: the monitoring questions, review cadence, and revision log are built into the system that collects the data, so the plan runs itself instead of sitting on a shelf.
MEL runs on the theory of change: the outcome stages become the indicators monitoring tracks, and the assumptions named under each causal arrow become the questions the quarterly learning review checks. That connection is what turns a falling number into a decision, because the assumption tells you which link failed, not just that an outcome was low. Sopact ties every monitoring signal to a theory-of-change assumption so the learning review acts on the exact link that broke rather than on a vague trend.
Because learning requires evidence to return in time to change something, and traditional evaluation runs on an annual cadence slower than the program itself, so by the time the report lands the cohort it describes has graduated. The fix is timing, not intention: monitor assumptions mid-cycle, schedule a review often enough to act, and document every revision. Sopact builds all three into the system, so the learning loop closes on a quarterly cadence instead of being promised in a proposal and forgotten.