MEL Basics 101 | ImpactMojo

ImpactMojoMEL Basics 101www.impactmojo.in

ImpactMojo 101 Series · Free Forever

MEL
Basics
101

Monitoring, Evaluation & Learning — A Practical Introduction for Development Practitioners

Monitoring Evaluation Learning 100 Slides Free Access

ImpactMojoMEL Basics 101www.impactmojo.in

Agenda

What We Cover in 100 Slides

What Is MEL and Why It Matters

Slides 3–10

Theory of Change

Slides 11–18

Results Frameworks & Logframes

Slides 19–26

Indicators: Designing & Selecting

Slides 27–34

Data Collection Methods

Slides 35–43

Evaluation Design

Slides 44–52

Attribution & Contribution

Slides 53–60

Learning Systems & Adaptive Management

Slides 61–68

Communicating Evidence

Slides 69–76

MEL in Practice: Common Failures

Slides 77–84

MEL for Different Contexts

Slides 85–92

Further Reading & Resources

Slides 93–100

ImpactMojoMEL Basics 101www.impactmojo.in

Section One

What Is MEL and Why It Matters

ImpactMojoMEL Basics 101www.impactmojo.in

Section 01 · Foundations

MEL: What Each Letter Actually Means

M — Monitoring

Tracking What Is Happening

Continuous, systematic collection of data on programme activities and outputs. Tells you what was done and whether it was done as planned. Ongoing throughout programme life. Primarily for management.

E — Evaluation

Assessing What It Achieved

Periodic, structured assessment of programme effectiveness, efficiency, relevance, and impact. Asks whether the programme worked, for whom, and why. Conducted at defined points — midterm and/or end of programme.

L — Learning

Using Evidence to Improve

The processes by which evidence from monitoring and evaluation feeds back into decision-making to improve programmes and strategy. Often the most underdeveloped of the three. Knowledge management, adaptation, and institutional memory.

Why the order matters: Monitoring without evaluation tells you what happened but not whether it mattered. Evaluation without monitoring lacks reliable data. Both without learning produce reports nobody acts on. MEL as an integrated system — not three separate functions — is the key shift from compliance to genuine improvement.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 01 · Why MEL Matters

Why MEL Matters: Four Distinct Functions

Accountability

Demonstrating to funders, government, and communities that resources were used as intended and achieved stated results. Required for donor compliance. Also — accountability to beneficiaries.

Learning

Understanding what works, for whom, and in what conditions — to improve current and future programming. The least prioritised function but the most valuable long-term.

Management

Real-time information for operational decisions — is implementation on track? Are there early warning signs of problems? Allows course correction before it's too late.

Advocacy & Influence

Evidence from well-designed evaluations can influence policy, funding priorities, and sector practice beyond the organisation. Impact at scale through knowledge sharing.

The accountability-learning tension: MEL systems designed primarily for donor accountability tend to produce data that looks good on paper but tells organisations little about what is actually working. Organisations prioritise what gets measured for compliance — not what generates genuine learning. The challenge is designing MEL systems that serve both functions without one colonising the other.

$20B+

Spent annually on development evaluation globally — most in compliance reporting rather than learning. 3ie estimates.

40%

Of major NGO evaluations are rated as not used to improve programming — Oxfam internal review 2019

ImpactMojoMEL Basics 101www.impactmojo.in

Section 01 · History

A Brief History of Development Evaluation

Period	Dominant Approach	Key Features
1950s–70s	Project evaluation	Focus on infrastructure; input-output accounting; World Bank led; economic rates of return
1970s–80s	Logical Framework introduced	USAID develops logframe (1969); structured planning tool becomes sector standard; criticism begins immediately
1980s–90s	Participatory methods	PRA (Chambers), PAR emerge; community voice in assessment; qualitative methods gain legitimacy
1990s–2000s	Results-based management	New Public Management; donor pressure for outcomes not inputs; MDGs accelerate indicator proliferation
2000s–10s	RCT revolution	Randomised trials as "gold standard"; Banerjee, Duflo, Kremer; J-PAL; 3ie; controversy about external validity
2010s–present	Learning and complexity	Developmental evaluation; PDIA; adaptive management; systems thinking; MEL as culture not compliance

Why history matters for practice: The tools and frameworks currently used in MEL carry the assumptions of their origins. Logframes were designed for discrete infrastructure projects — not complex, adaptive social change programmes. Understanding where a tool came from helps practitioners know when it fits and when to look elsewhere.

The Indian context: India's development sector MEL is shaped by multiple overlapping traditions: government MIS systems (NITI Aayog, DMEO), donor-driven logframe and RBM requirements, academic evaluation culture (NIRD, IIPA), and increasingly CSO-led participatory approaches. No single approach dominates — practitioners navigate all of them simultaneously.

DMEO: Development Monitoring and Evaluation Office under NITI Aayog — India's central government evaluation body. Evaluates flagship government schemes (MGNREGS, PMGSY, Aspirational Districts Programme). Its reports are often more candid than ministry self-evaluations and are publicly available.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 01 · MEL Landscape in India

MEL in India: Who Demands It and What They Want

India's MEL landscape is shaped by who is funding and who is implementing. Each principal in the funding chain has different MEL requirements — which are often not aligned with each other and sometimes not with programme improvement either.

Principal	What They Want from MEL
International donors (FCDO, USAID, EU)	Standardised indicators; logframe compliance; outcomes data; value for money; gender disaggregation
Indian philanthropies (TATA, Azim Premji, Gates)	Theory-based evaluations; learning focus; innovation evidence; scale-readiness assessment
CSR funders	Output counts for reporting; beneficiary stories; visibility; annual reporting cycle
Government (MoU partners, state govt)	Alignment with government schemes; convergence data; MIS integration; political sensitivity
Communities	Accountability to them; participation in assessment; feedback on what isn't working; recognition

The multiple principal problem: An organisation funded by FCDO, an Indian foundation, and CSR simultaneously may face three different indicator frameworks, three different reporting templates, and three different evaluation timelines — all for the same programme. MEL harmonisation across funders is one of the most practically significant sector challenges.

FCRA and MEL: Post-2020 FCRA amendments have reduced international funding flows to Indian CSOs. This has shifted MEL demands — more domestic funders (Indian foundations, CSR) who have different and often less standardised MEL requirements. Some CSOs report MEL burden decreasing with shift to domestic funders; others find domestic funder requirements less rigorous but less useful for learning.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 01 · Core Concepts

The Results Chain: Inputs, Activities, Outputs, Outcomes, Impact

Inputs

→

Activities

→

Outputs

→

Outcomes

→

Impact

Inputs

Money, staff, materials, time, expertise invested in the programme

Example: ₹50L budget, 3 field staff, training materials

Activities

What the programme does with those inputs — the work

Example: Training sessions, farmer group meetings, seed distribution

Outputs

Direct, tangible products of activities. Within programme control.

Example: 500 farmers trained, 200 farmer groups formed

Outcomes

Changes in knowledge, behaviour, practice, or conditions. Medium-term.

Example: Farmers adopt climate-resilient varieties; yields stable

Impact

Long-term, fundamental change in people's lives. Often outside programme control alone.

Example: Reduced rural poverty; improved food security over 5+ years

The monitoring-evaluation divide: Monitoring typically tracks inputs, activities, and outputs — what is within programme control and measurable in real time. Evaluation assesses outcomes and impact — changes in the world that require longer time horizons and more rigorous methods. This is why evaluation cannot replace monitoring, and vice versa.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 01 · Outputs vs Outcomes

The Output Trap: Why Counting Activities Is Not Measuring Change

The most common MEL failure in development practice is conflating outputs with outcomes — reporting on what was done as if it demonstrates what changed. Outputs are necessary but not sufficient. A training session (output) does not automatically produce changed practice (outcome). The gap between them is where programmes frequently fail.

The output trap in action: An NGO reports: "We conducted 240 training sessions reaching 4,800 participants." This is an output. The questions it cannot answer: Did participants learn? Did they change their behaviour? Did their situation improve? Did the change last? Without outcome data, the training count proves only that activities happened — not that they mattered.

Why organisations fall into this trap: Outputs are easy to count, available quickly, and within the organisation's direct control. Outcomes require time, methodological effort, and often show mixed results that are hard to present positively. Donor reporting requirements often accept output data, removing the incentive to measure harder.

Output vs Outcome: Recognising the Difference

Programme Area	Output (what happened)	Outcome (what changed)
Education	500 children enrolled in school	500 children reading at grade level by age 10
Health	10,000 bed nets distributed	Malaria incidence reduced 30% in target area
Livelihoods	200 women received seed capital	200 women with sustainable income above poverty line after 2 years
WASH	500 toilets constructed	500 households using toilets consistently (ODF status maintained)
Agriculture	300 farmers trained on SRI	250 farmers practising SRI with documented yield improvement

The test: Ask "so what?" after each indicator. "We trained 300 farmers" — so what? "Farmers' yields improved 20%" — so what? "Household food security improved" — so what? Keep asking until you reach an answer that is about fundamental change in people's lives. That is your impact level.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 01 · MEL as a System

MEL as an Integrated System: The Architecture

Effective MEL is not a set of tools used at different times — it is an integrated system where each component informs and depends on the others. The system starts with a clear Theory of Change and flows through all programme phases.

MEL System Architecture

Theory of Change

Articulates the change hypothesis — what are you trying to achieve and how?

Results Framework / Logframe

Operationalises ToC into measurable indicators across the results chain

MEL Plan

Specifies who collects what data, how, and when — the operational blueprint

Data Collection & Management

Surveys, interviews, observations, records — with quality control and storage

Analysis, Reporting & Learning

Making sense of data and feeding it back into programme decisions

The MEL plan: One of the most neglected documents in development programming is the MEL Plan — a detailed, operational document specifying every indicator, its definition, its data source, the collection tool, the responsible person, the collection frequency, and the analysis method. Without a MEL plan, monitoring becomes ad hoc and evaluation becomes a scramble.

Design phase

MEL system must be designed before the programme starts — not added on after problems emerge. Retrofit MEL is consistently inferior.

5–10%

Industry norm for MEL as share of programme budget. Underfunded MEL produces incomplete data. Overfunded MEL produces data nobody uses.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Two

Theory of Change

ImpactMojoMEL Basics 101www.impactmojo.in

Section 02 · Theory of Change

Theory of Change: The Programme's Causal Hypothesis

Theory of Change

An explicit articulation of how and why a programme expects its activities to lead to intended outcomes and impact. It maps the causal pathway from inputs to change — specifying the assumptions that must hold for each step to follow from the previous one. A ToC is a testable hypothesis about how social change happens.

"If-then-because" logic: A ToC can be expressed as a series of if-then statements: "If we provide climate-resilient seeds (activity), then farmers will adopt them (outcome) — because farmers are risk-averse and need reliable evidence of performance before changing practices (assumption)." The "because" — the assumption — is where most programmes fail to be explicit.

What a Good ToC Includes

Problem analysis: Clear articulation of the problem being addressed and its root causes — not just symptoms
Change hypothesis: How the intervention will address those causes — the mechanism of change
Causal pathways: The steps between activities and impact, with each link made explicit
Assumptions: The conditions that must hold at each step for the logic to work — external factors, enabling conditions, community readiness
Boundaries: What the programme will and won't address — contextual factors it works with but cannot change
Stakeholders: Whose behaviour needs to change; who enables or blocks the change
Preconditions: What must already exist for the programme to work as designed

ImpactMojoMEL Basics 101www.impactmojo.in

Section 02 · Assumptions

Assumptions: The Hidden Infrastructure of a Theory of Change

Assumptions are the conditions that must hold for the causal logic of a ToC to work. They are not risks — they are the unstated premises in the programme's reasoning. Making them explicit is one of the most valuable exercises a programme team can do, because untested assumptions are the most common cause of programme failure.

The assumption audit: For each arrow in a results chain, ask: "What must be true for this step to follow from the previous one?" Write it down. Then ask: "Do we have evidence this is true in our context? Or are we assuming it?" Untested assumptions about beneficiary behaviour, community readiness, or enabling environment are where most ToCs unravel.

Classic failing assumptions in Indian development: "Women will use the savings account once it is opened" — assumption: they have control over household finances. "Farmers will adopt the improved seed" — assumption: they can access credit to buy it. "Communities will maintain the water pump" — assumption: a functioning maintenance committee will form. Each is an untested social assumption, not a logistical one.

Assumption Types and How to Test Them

Assumption Type	Example	How to Test
Behavioural	Participants will attend training if offered	Pilot; attendance data; FGDs on barriers
Contextual	Government services are accessible and functional	Baseline mapping; key informant interviews
Causal	Knowledge change leads to behaviour change	Literature review; tracer studies post-training
Social norm	Husbands will support wives' participation	Formative research; KAP surveys; FGDs
Market	Buyers will pay premium for quality produce	Market research; pilot sales; price surveys

ImpactMojoMEL Basics 101www.impactmojo.in

Section 02 · Building a ToC

Building a Theory of Change: Process Matters as Much as Product

A ToC developed by one person on a laptop for a funding proposal has little value. The process of building a ToC — who is in the room, what evidence informs it, whether it gets challenged and revised — determines whether it becomes a genuine guide to practice or a compliance document.

Start with the problem, not the solution

Problem tree analysis; root cause identification; whose problem is it?

Work backwards from long-term impact

What does success look like in 10 years? What needs to happen for that?

Involve programme staff, community, and partners

Participatory ToC workshops surface hidden assumptions and knowledge

Stress-test every causal link

Ask "why would this lead to that?" for every arrow; name the assumption; assess evidence

Revise regularly — not just at design

ToC is a living document; update when assumptions are violated or context changes

ToC vs logframe: A Theory of Change comes before a logframe. The ToC articulates the change hypothesis in narrative and visual form — including complexity, assumptions, and contested claims. The logframe then translates the ToC into a structured planning matrix with measurable indicators. Many organisations skip the ToC and go straight to logframe — producing a plan that nobody can explain in plain language.

Common ToC weaknesses in Indian NGO proposals: (1) The ToC is a post-hoc justification for a programme already designed rather than a design tool; (2) Activities are listed as outcomes ("200 women will form SHGs" — that is an output, not a change); (3) The causal pathway from community to systemic change is missing — the ToC stops at individual behaviour change and never explains how that aggregates to structural impact; (4) Assumptions are not named at all.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 02 · Complexity

Complexity and Theory of Change: When Simple Logic Fails

The standard results chain assumes linear causality — activities lead predictably to outputs lead to outcomes lead to impact. But many development challenges are complex — meaning they involve non-linear dynamics, emergent outcomes, and multiple feedback loops that a linear ToC cannot capture.

Complicated vs Complex

Complicated problems have many parts but are ultimately knowable — like building a bridge. Best practice can be identified and replicated. Complex problems involve human systems, power, culture, and feedback loops — like reducing domestic violence or improving governance. You can't reliably predict outcomes. The Cynefin framework distinguishes these.

Implications for MEL

Emergent outcomes: Complex programmes produce outcomes that were not anticipated in the original ToC — positive and negative. MEL systems must be able to capture and learn from emergence, not just measure pre-specified indicators.
Multiple pathways: In complex systems, the same input can produce different outcomes in different contexts. MEL must track context, not just intervention, to explain variation.
Feedback loops: Outcomes loop back to change the conditions for the programme — positive feedback can amplify success; negative feedback can undermine it. A static logframe cannot capture this dynamic.
Contribution not attribution: In complex systems, you rarely caused an outcome alone. The appropriate question shifts from "did we cause this?" to "what was our contribution?" — requiring different evaluation methods.

Practical response: For complex programmes, use a ToC that explicitly maps feedback loops, names multiple pathways, and builds in regular revision cycles. PDIA (Problem-Driven Iterative Adaptation) and developmental evaluation are specifically designed for complex change — not logframe-driven planning.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 02 · ToC Examples

Two ToC Examples: Weak vs Strong

Weak ToC: Girls' Education Programme

"We will build schools and train teachers. Girls will attend school, learn, and this will increase women's economic participation and reduce poverty."

✗No explanation of why girls aren't in school now
✗No mention of household-level barriers (marriage pressure, safety)
✗Massive leap from "attend school" to "economic participation"
✗No assumptions stated
✗No explanation of which girls, which pathway, what timeline

Stronger ToC: Girls' Education Programme

"Girls in our target area face dual barriers: household-level pressure for early marriage and inadequate menstrual hygiene facilities in schools. We address both: by working with parents and community leaders to shift norms around girls' education (assumption: norm-shift is possible in 3 years), and by upgrading school WASH facilities (assumption: infrastructure is a binding constraint). If norms shift AND infrastructure improves, enrolment and retention will increase — leading, over 5–10 years, to higher women's education levels, which evidence shows is associated with delayed marriage, lower fertility, and improved household incomes. We acknowledge that employment opportunities and social norms beyond school also determine economic outcomes — these are outside our direct control."

✓Problem analysis explicit
✓Dual pathway explained
✓Assumptions named and testable
✓Boundaries of programme acknowledged
✓Evidence base cited

ImpactMojoMEL Basics 101www.impactmojo.in

Section 02 · Change Mechanisms

Change Mechanisms: How Development Programmes Produce Change

Behind every ToC is an implicit theory of how change happens. Making this explicit — the mechanism of change — is what separates a logistically coherent plan from a theoretically grounded one. Different mechanisms require different interventions, different timelines, and different evidence bases.

Mechanism	How It Works	Example
Information / KAP	Knowledge → attitude → practice: people change behaviour when they know better. Weak for entrenched behaviour.	Health education campaigns; nutrition counselling
Economic incentives	Behaviour changes when it is more profitable or cheaper. Effective for market-oriented change.	Conditional cash transfers; insurance products; market linkages
Norm change	Social norms drive behaviour. Change requires critical mass shifts in what is seen as acceptable.	Community conversation models; SASA!; Tostan
Power shift	Marginalised groups gain power (economic, political, social) to claim rights and resources.	SHG federations; collective bargaining; rights-based advocacy
Systems change	Changes in rules, policies, and institutions create enabling environment for improved outcomes.	Policy advocacy; institutional strengthening; market systems development

The mechanism shapes the measurement: If your mechanism is "information leads to knowledge leads to behaviour change" — then you need to measure knowledge change AND behaviour change AND test whether the knowledge actually drove the behaviour (not just whether both changed). If your mechanism is "economic incentive" — you need price data, adoption data, and an understanding of the price point at which behaviour shifts.

Mechanism mismatch: One of the most common ToC failures is implementing an information-based intervention for a behaviour driven by social norms or economic constraints. Training women on maternal nutrition does not improve nutritional outcomes if the binding constraint is household food insecurity or women's lack of control over household food decisions. The mechanism doesn't match the problem.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 02 · Synthesis

Theory of Change: Quality Checklist

Signs of a Strong ToC

Problem analysis is evidence-based and based on consultation with affected communities
Causal pathways are explicit — each arrow can be explained in a sentence
Assumptions are named, not buried
The mechanism of change is specified and matches the evidence base
Long-term impact is distinguished from near-term outcomes
Boundaries of the programme are acknowledged — what it won't change
Multiple stakeholders' roles are identified
The ToC can be explained clearly to community members

Red Flags in a Weak ToC

Outputs are labelled as outcomes ("200 women trained" as a change outcome)
The ToC is a diagram with boxes and arrows but no explanatory text
Assumptions section is blank or says "no major assumptions"
The pathway from individual change to social/systemic change is missing
No time horizon specified for outcomes vs impact
The same ToC applies regardless of context — no adaptation for different geographies
Nobody on the programme team can explain it in plain language

The living ToC: A ToC should be revisited at least annually — not treated as a fixed document. When monitoring data shows that a step in the chain isn't working as expected, the ToC either needs to be updated (the theory was wrong) or the programme design needs to change (the theory is right but the implementation isn't achieving the prerequisite conditions).

ImpactMojoMEL Basics 101www.impactmojo.in

Section Three

Results Frameworks & Logframes

ImpactMojoMEL Basics 101www.impactmojo.in

Section 03 · Logframe

The Logical Framework: Structure and How to Read It

The Logframe (Logical Framework Matrix) is a 4×4 planning and monitoring tool developed by USAID in 1969. It remains the most widely used planning format in international development — and the most widely misused. Understanding its structure is essential before knowing its limits.

Hierarchy	Indicators	Means of Verification	Assumptions
Goal / Impact Long-term change	How we know goal is achieved	Data sources for goal indicators	Conditions beyond programme for goal to follow from purpose
Purpose / Outcome Why we're doing this	How we know purpose is achieved	Data sources for purpose indicators	Conditions beyond programme for purpose to contribute to goal
Outputs What we produce	How we know outputs are achieved	Data sources for output indicators	Conditions beyond programme for outputs to lead to purpose
Activities What we do	Inputs required	Budget lines	Conditions beyond programme for activities to produce outputs

The "if-and" logic: The logframe has vertical logic (if you do the activities AND assumptions hold, outputs are produced; if outputs are produced AND assumptions hold, purpose is achieved) and horizontal logic (purpose indicators with means of verification in the same row). The assumptions column is where the logic becomes testable — and where most logframes go wrong by leaving it blank.

Classic logframe errors: (1) Activities listed as outputs; (2) Outputs listed as outcomes; (3) Assumptions column blank or saying "favourable policy environment" without specifics; (4) Indicators with no means of verification; (5) Same indicator at multiple levels; (6) More than 5–7 indicators per level (indicator proliferation); (7) The logframe bears no relationship to the programme ToC.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 03 · Results Framework

Results Frameworks: Beyond the Logframe

A Results Framework is a broader term for any structured representation of expected programme results. Different donors and organisations use different formats. Understanding the common ones allows practitioners to translate between them and choose the right tool for the context.

Framework	Used By	Key Feature
Logframe	EU, DFID/FCDO (older), most bilateral donors	4×4 matrix; vertical and horizontal logic; assumptions column
Results Framework (RF)	USAID	Hierarchical results pyramid; strategic objectives, intermediate results; performance management plan
Performance Framework	USAID/PEPFAR	Indicator-focused; progress vs target tracking; less causal theory
Impact Map	Social enterprise; SROI	Maps from activities to outputs to outcomes to impacts; used for Social Return on Investment
Outcome Harvesting	Complex, systems programmes	Identifies outcomes that have occurred; works backwards to contribution; no pre-specified indicators

USAID vs FCDO logics: USAID's results framework separates strategic objectives (SOs) from intermediate results (IRs) in a tree structure rather than a matrix. The logic is similar — but the terminology and visual format differ. FCDO's newer programme approaches often use Theory of Change diagrams with Outcome Harvesting rather than traditional logframes. Most practitioners need to be fluent in at least two of these systems.

Indicator fatigue: Large development programmes — particularly government or multi-donor — can end up with 50–100 indicators across a results framework. This is unmanageable. The purpose of the framework is defeated — teams spend all their time collecting data for compliance reporting and have no time to analyse or learn. The discipline of limiting indicators to the essential is one of the most important MEL skills.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 03 · Workplans & Milestones

From Framework to Implementation: Workplans and Milestones

A results framework tells you what to achieve. A workplan tells you how and when. Milestones are the intermediate checkpoints that indicate whether the programme is on track. Good workplanning is the bridge between strategic planning and operational reality.

Annual workplan vs MEL plan: These are different documents that must be aligned. The workplan specifies activities, responsible staff, timelines, and budgets. The MEL plan specifies what data will be collected, when, by whom, using what tools, and how it will be analysed. When they are developed separately — as often happens — the MEL plan ends up collecting data nobody uses and missing data the workplan needs.

Milestone vs target: A milestone is a qualitative marker of progress — "agricultural groups formed and functional" by month 6. A target is a quantitative level for an indicator — "200 farmers trained" by month 6. Both are needed. Programmes that only set targets can meet their numbers while missing the qualitative preconditions for outcome achievement.

A Realistic Workplan: What It Includes

Activity-level detail: Sub-activities, not just activity names. "Training" is not enough — specify preparation, participant recruitment, delivery, follow-up, and data collection from training
Resource linkages: Each activity linked to budget line and responsible staff — not just a name but a designated person who knows they are responsible
Realistic timelines: Built with input from field staff, not by headquarters. Field realities (monsoon season, harvest period, festival calendar) must be factored in
MEL data collection events: Baseline survey, midline, endline, FGDs, case studies — all in the workplan with timelines and budgets, not afterthoughts
Review and adaptation points: Quarterly review meetings, progress against milestones, documented decisions — not just monthly financial reporting

ImpactMojoMEL Basics 101www.impactmojo.in

Section 03 · Logframe Critique

Critiquing the Logframe: Why It Doesn't Fit All Contexts

The logframe has been criticised since its introduction. Its dominance in development reflects donor power rather than technical superiority — organisations use it because funders require it, not because it is the best tool for complex social change. Understanding the critique helps practitioners use it better and know when to supplement it.

Linear causality assumed: Logframes assume activities lead predictably to outcomes. Real development is non-linear, context-dependent, and emergent.
Designed for infrastructure projects: Originally designed for discrete, bounded construction projects — not for behaviour change, governance, or systems work.
Assumes pre-specified outcomes: Social change programmes often don't know what outcomes will emerge. Requiring pre-specified indicators creates perverse incentives to only measure what was planned.
Inhibits adaptation: Once a logframe is agreed with a donor, changing it requires formal amendment processes — creating institutional resistance to adaptive management.
Concentrates power: Logframes are typically designed by external consultants and headquartered staff, not communities. They embed external assumptions about what change should look like.

When Logframes Work and When They Don't

Logframes work well for

Service delivery programmes; well-understood causal pathways; discrete outputs; large-scale programmes needing standardised data; multi-country programmes requiring comparability

Logframes work poorly for

Advocacy and governance; complex adaptive programmes; community-led initiatives; systems change; innovation programmes; programmes in fragile/conflict-affected contexts

The hybrid approach: Many experienced practitioners use a logframe for donor compliance while running a parallel, richer MEL system (with outcome harvesting, most significant change, or developmental evaluation) for genuine learning. The logframe satisfies the funder; the richer system informs the team. This dual-track approach is pragmatic but resource-intensive.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 03 · Results Matrix

The Results Matrix: A More Flexible Planning Tool

For programmes that find the logframe too rigid, the Results Matrix (or Results Measurement Framework) offers a more flexible structure — keeping the discipline of a results chain while allowing more nuance in how outcomes are described and measured. It is not a replacement for a ToC but an operationalisation of it.

Result Level	Statement	Indicator	Baseline	Target	Data Source	Timeline
Impact	Reduced rural poverty in target districts	% HHs below poverty line	42%	32%	Household survey	Year 5
Outcome	Improved agricultural income	Avg annual HH farm income (₹)	₹42,000	₹58,000	Household survey	Year 3
Output	Farmers adopt climate-resilient varieties	% farmers using improved seed	18%	60%	Adoption survey	Year 2
Activity	Training and demo plots established	No. of demo plots operational	0	50	Field register	Year 1

The baseline column: One of the most commonly neglected elements in results matrices. Without a baseline — a measurement of the indicator's starting value before the programme — you cannot know whether change occurred during the programme. Baseline data collection must be planned before programme start and budgeted accordingly. Post-hoc baseline reconstruction is methodologically unreliable.

Target-setting discipline: Targets must be based on evidence and realistic assessment — not aspirational numbers chosen to look ambitious in a proposal. Targets that are met too easily signal they were set too low. Targets never met signal they were unrealistic or the programme model was wrong. Both undermine credibility and learning. Consult field staff, review comparable programme data, and build in target revision at midterm.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 03 · Baseline

Baselines: Why They Are Non-Negotiable

Baseline

A measurement of outcome and impact indicators taken before the programme begins (or at the earliest feasible point). Establishes the starting value against which change will be measured. Without a baseline, you can measure the endpoint but not the change — so you cannot know whether the programme caused the measured condition or inherited it.

Before

Baseline must be collected BEFORE the programme begins. Collecting it 6 months in means the programme has already started changing things.

Same method

Baseline and endline must use identical instruments, sampling methods, and data collection protocols — or the comparison is invalid.

Baseline Design Considerations

What to measure: Baseline should cover all outcome-level indicators in the results framework — not just outputs. Many programmes baseline only on output-level variables and then cannot measure outcome change.
Comparison population: If the evaluation design involves a comparison group (non-beneficiaries), baseline must be collected from both groups simultaneously. Do not baseline the control group later — conditions change.
Disaggregation: Baseline data must be collected with the same disaggregation planned for endline — gender, age, caste, geography. If the endline will report on women's outcomes separately, the baseline needs women's data separately.
Sample size: Baseline sample must be large enough to detect the expected change at endline with statistical power. Under-powered baselines lead to endlines that show no significant change even when real change occurred.
Administrative data: Where primary data collection is not possible, administrative data (government MIS, health records, school enrollment data) can provide baseline — but quality must be assessed and limitations documented.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 03 · Synthesis

Results Frameworks: Practical Checklist

Element	Quality Standard	Common Failure
Theory of Change	Explicit, evidence-based, tested with field team and community	Missing entirely; or logframe passed off as ToC
Levels of results	Clear distinction between activity, output, outcome, impact with appropriate indicators at each	Activities mislabelled as outputs; outputs as outcomes
Indicators	5–7 per level maximum; SMART; gender-disaggregated; based on evidence	50+ indicators; unmeasurable; activity counts at outcome level
Baseline	Collected before programme start; same sampling and tools as endline; disaggregated	Not collected; collected late; not disaggregated
Targets	Evidence-based; realistic; midterm and endline; includes decline scenarios	Aspirational; no midterm; no process for revision
Means of verification	Specific, accessible, reliable data source named for each indicator	"Field reports" or "project records" — not specific enough
Assumptions	Named; testable; with monitoring plan for critical assumptions	Blank; or generic ("stable political context")

The living framework: A results framework designed at proposal stage must be treated as provisional. Baseline findings, implementation realities, and context changes should trigger a formal framework review. Building in a midterm review process is the most important institutional safeguard against frameworks that become irrelevant to the programmes they describe.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Four

Indicators: Designing & Selecting

ImpactMojoMEL Basics 101www.impactmojo.in

Section 04 · Indicators

What Makes a Good Indicator

Indicator

A variable that provides a measurable representation of a result. It tells you what to observe to know whether a result has been achieved. An indicator is not the result itself — it is a proxy for it. The quality of your measurement is only as good as the quality of your indicators.

SMART Indicators

Specific

Precisely defined — what is being measured, in whom, where, when

Measurable

Can be observed and quantified (or verified qualitatively) reliably

Achievable

The target level is realistic given programme scope and timeline

Relevant

Directly linked to the result it is meant to measure — not a proxy for a proxy

Time-bound

Measured at a specific point in time or over a defined period

Indicator Anatomy: What Must Be Specified

Indicator name: "% of women with access to credit" — this is not specific enough. Does "access" mean applied? Approved? Disbursed? Used?
Definition: Precise operational definition of each term in the indicator. "Women" — all women in the target area? Only those enrolled in the programme?
Numerator and denominator: For percentage indicators — exactly what goes in each. "% of farmers practising SRI" = farmers practising SRI / farmers receiving SRI training × 100
Disaggregation: By gender, age, caste/ethnicity, geography — specified in advance, not chosen after data is collected
Data source: Specific named source. "Household survey" — who collects it, how often, with what instrument?
Baseline value and target: Numeric values with source and date

The indicator reference sheet: Every indicator in a results framework should have a one-page reference sheet with all of the above. Without it, different field staff will collect the same indicator differently — making aggregation and comparison impossible.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 04 · Indicator Types

Types of Indicators: Quantitative, Qualitative, and Proxy

Type	What It Measures	When to Use	Limitation
Quantitative	Numeric counts, percentages, averages, rates	Outputs, service coverage, income, assets	Doesn't capture quality, process, or meaning
Qualitative	Experiences, perceptions, processes, meanings	Behaviour change, empowerment, satisfaction, norms	Harder to aggregate; requires skilled collectors; more expensive
Proxy	An observable variable used to represent a harder-to-measure concept	When direct measurement is not feasible — e.g. "nights slept under net" as proxy for malaria prevention behaviour	Proxy may not actually reflect the underlying concept accurately
Process	Quality and fidelity of implementation	Understanding why outcomes did or didn't occur; intervention fidelity	Easy to mistake for outputs; needs clear quality standards
Sentinel	A few key indicators tracked continuously for early warning	Rapid feedback loops; detecting programme drift early	May miss slow-moving changes not captured by sentinel points

The mixed-methods approach: Quantitative indicators tell you how much changed; qualitative indicators tell you why, for whom, and under what conditions. A programme that only tracks quantitative indicators can miss systematic exclusion of sub-groups (who show up in aggregate numbers but not in disaggregated analysis) and can fail to understand why an outcome did or didn't occur. Most outcome-level results frameworks benefit from including at least one qualitative indicator per outcome.

Proxy indicator caution: A classic case in health: "number of ANC visits" as a proxy for improved maternal health outcomes. ANC visits are easier to count than outcomes — but the link is not automatic. If ANC services are of poor quality, attendance does not produce health benefits. Always ask: "Does this proxy actually lead to the outcome it's supposed to represent?"

ImpactMojoMEL Basics 101www.impactmojo.in

Section 04 · Gender in MEL

Gender-Responsive MEL: Disaggregation Is Not Enough

Gender-responsive MEL requires more than disaggregating data by sex. It requires asking different questions about different experiences, using methods that reach women, and measuring changes in gender relations and power — not just women's participation in activities.

The disaggregation fallacy: Disaggregating "% of participants who are women" is gender-sensitive reporting. Gender-responsive MEL asks: Did women and men experience the programme differently? Did women have equal access? Did the programme change women's decision-making power — or just their participation in a programme? Did women's outcomes improve at the expense of their time and unpaid care burden?

The Women's Empowerment in Agriculture Index (WEAI): Developed by IFPRI and used widely in agricultural programmes. Measures 5 domains: decisions about agriculture; access to resources; control of income; leadership; time allocation. Standard instrument available free — appropriate for smallholder agriculture programmes in India. Captures nuances of intra-household dynamics that simple "% women trained" cannot.

Gender-Responsive Indicator Examples

Standard Indicator	Gender-Responsive Version
% of HHs with savings	% of women with savings in their own name with independent access
% of farmers adopted improved seed	% of female-headed HHs AND % of women in male-headed HHs with own decision-making on seed choice
% of children completing school	% by gender; AND reasons for dropout disaggregated by gender
HH income increased	Women's control over HH income; women's independent income; intra-HH resource allocation
Community participation in planning	Women's meaningful participation (not just attendance); women's proposals adopted in community plans

ImpactMojoMEL Basics 101www.impactmojo.in

Section 04 · Selecting Indicators

Selecting Indicators: The Discipline of Less

Indicator selection is a discipline of subtraction. The natural tendency is to add — "let's also measure X" — without considering the data collection burden, analysis capacity, or whether X actually informs decisions. Excellent MEL systems are notable for what they choose not to measure.

The 5 test questions for any proposed indicator:
(1) Is this indicator directly linked to a result in our framework?
(2) Do we have the capacity to collect this data reliably?
(3) Will this data actually be used to make a decision?
(4) Do we already have this data from another source?
(5) What would we do differently if this indicator showed good results vs poor results?
If you can't answer all five — especially (3) and (5) — cut the indicator.

Indicator Selection Criteria

What makes organisations select indicators (survey of 200 Indian NGOs)

Source: ImpactMojo illustrative data based on sector experience — not a published study

ImpactMojoMEL Basics 101www.impactmojo.in

Section 04 · Standard Indicators

Standard vs Custom Indicators: When to Use Each

Standard indicators — like those from USAID's Foreign Assistance Indicator Library, WHO standard definitions, or India's HMIS — have validated definitions, established baselines, and allow comparison across programmes and geographies. Custom indicators are specific to a programme's unique theory of change. Both have a place.

Standard Indicators: Use When

Sector benchmarking is needed; donor requires them; comparable data exists; results will be used to influence sector policy; they align with your ToC outcomes

Custom Indicators: Use When

Your ToC targets a unique change pathway; standard indicators don't capture your mechanism; local context requires adapted definitions; you need to learn about something not in standard libraries

Commonly Used Standard Indicators in India Development

Domain	Standard Indicator	Source
Health	Institutional delivery rate; ANC visits ≥4; stunting prevalence (HAZ <−2)	NFHS; HMIS
Education	Gross/Net Enrollment Ratio; foundational literacy rate; dropout rate	UDISE+; ASER
Poverty	Consumption per capita (MPI); multidimensional poverty index (OPHI)	NSSO; NFHS
Agriculture	Crop yield (kg/acre); cropping intensity; area under irrigation	State Agri Dept; ICRISAT
WASH	HHs with piped water; ODF status; HHs with handwashing facility with soap	NFHS; SBM MIS

ImpactMojoMEL Basics 101www.impactmojo.in

Section 04 · Indicator Problems

When Indicators Go Wrong: Goodhart's Law and Gaming

Goodhart's Law

"When a measure becomes a target, it ceases to be a good measure." Once an indicator is used for accountability (reporting to funders, evaluating staff performance), people optimise for the indicator — not for the underlying result the indicator was supposed to represent. The measure loses its signal value.

Classic development examples: (1) "Number of trainings conducted" as a target → teams conduct shorter, lower-quality trainings to meet numbers. (2) "% of children enrolled in school" → enrollments inflated by registering children who don't attend. (3) "# of community leaders engaged" → meetings held with cooperative leaders rather than marginalised voices. In each case, the indicator target was met; the underlying objective was not.

How to Reduce Gaming Risk

Use indicator combinations: Track both enrolment AND attendance AND learning outcomes — gaming one of three is harder than gaming one of one
Include quality alongside quantity: "Number trained" paired with "% who demonstrate competency in post-training assessment" closes the gaming gap
Unannounced verification: Random spot-checks, beneficiary verification surveys, and third-party verification reduce data fabrication
Protect the data collection function: Field staff responsible for data collection should not be the same staff held accountable for the numbers — creates obvious conflict of interest
Make under-performance reportable without consequence: If reporting bad numbers leads to funding cuts or staff dismissal, teams will hide problems. Psychological safety for honest reporting is as important as indicator design

ImpactMojoMEL Basics 101www.impactmojo.in

Section 04 · Synthesis

Indicator Design: Quick Reference

Good Indicators Look Like This

"% of women in programme with independent bank account with at least ₹500 balance, disaggregated by marital status, measured at endline (18 months post-enrolment)"
"Average crop yield (kg/acre) for Kharif paddy, target beneficiary farmers, Year 2 post-training season, compared to Year 0 baseline"
"Community members (women and men, separately) who report feeling safe reporting violence to local authorities — % agree/strongly agree, endline survey"

Weak Indicators Look Like This

"Improved access to financial services" — what does access mean? For whom? Measured how?
"Agricultural productivity enhanced" — which crop? By how much? Compared to what?
"Community awareness raised" — what are they aware of? How is awareness verified?

Indicators by Result Level: Recommended Distribution

Source: ImpactMojo MEL guidance — industry norm for well-designed frameworks

The minimum viable indicator set: For a programme of 3–4 outputs with 2 outcomes, aim for: 2–3 impact indicators; 3–4 outcome indicators; 5–8 output indicators; 4–6 activity/process indicators. Total: 14–21. Any framework exceeding 30 indicators will likely not be used effectively. If your donor requires more, negotiate or build a parallel light-touch internal system.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Five

Data Collection Methods

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Surveys

Household Surveys: Design, Sampling, and Quality

Household surveys are the most common quantitative data collection method in development MEL. They are also one of the most error-prone — due to poor questionnaire design, inadequate sampling, untrained enumerators, and response bias. Understanding survey design basics is essential for anyone commissioning or interpreting survey-based MEL.

Questionnaire design principles: (1) Each question should measure one thing only; (2) Avoid leading questions; (3) Keep questions to a minimum — cognitive burden on respondents leads to lower quality answers over time; (4) Pilot-test with 10–20 respondents before full deployment; (5) Ensure cultural and linguistic appropriateness — especially in India's multilingual context.

Response bias in development MEL: Beneficiaries know they are being assessed by an organisation they depend on — and often give socially desirable answers. "Did the training help you?" asked by the same NGO that delivered it will produce overwhelmingly positive responses regardless of actual utility. Independent data collection, careful question framing, and anonymity assurance all reduce this bias — but never eliminate it entirely.

Sampling: The Fundamentals

Sampling Method	When to Use	Key Requirement
Simple Random Sampling	When complete list of population is available	Complete sampling frame (beneficiary list)
Systematic Random	Large lists; every nth element	Ordered list; no hidden patterns every n elements
Stratified Random	When subgroups (women, Adivasi, different districts) must be represented	Known stratum sizes; strata must be non-overlapping
Cluster Sampling	Geographically dispersed populations; efficiency	Clusters (villages) sampled, then households within; increases variance
Purposive	Qualitative studies; selecting information-rich cases	Clear criteria for selection; acknowledge non-representativeness

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Qualitative Methods

Qualitative Methods: FGDs, KIIs, and Observation

Qualitative methods generate depth, context, and explanation. They answer "why" and "how" — the mechanisms that quantitative surveys cannot reach. They are not less rigorous than quantitative methods — they are rigorous in different ways, requiring skilled facilitation, systematic analysis, and careful attention to positionality.

Key Qualitative Methods

Focus Group Discussion (FGD): Structured group discussion (6–8 participants) using a guide. Good for exploring shared norms, community-level perceptions, and generating hypotheses. Risk: dominant voices; social desirability in group settings.
Key Informant Interview (KII): In-depth interview with person with specific knowledge — health worker, teacher, community leader, government official. Semi-structured guide; typically 45–90 minutes.
In-depth Interview (IDI): Individual interview with programme participant. More personal than KII. Allows for probing experiences, perceptions, and behaviour change processes.
Observation: Direct observation of programme activities, community settings, or practices. Can validate or contradict self-reported data. Field notes; time-activity studies.

FGD vs IDI: when to use which: FGDs are efficient for generating community-level norms and shared perceptions. IDIs are appropriate for sensitive topics (domestic violence, income, family planning), individual experiences, and when social desirability in a group would suppress honest responses. Never use FGDs for sensitive or potentially shame-laden topics.

Conducting FGDs in rural India: practical issues: (1) Group composition matters — mixed caste groups in many contexts will produce dominant-caste narratives; separate by caste/community; (2) Gender-segregated groups almost always produce more honest data from women; (3) Introductory sessions about confidentiality and the purpose of the discussion significantly improve quality; (4) Co-facilitation (facilitator + note-taker) allows better observation and recording; (5) Avoid venue the programme runs — creates power dynamics that suppress criticism.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Participatory Methods

Participatory Methods: Communities as Evaluators

Participatory Rural Appraisal (PRA) and its variants give communities the tools to assess their own situations and the changes they experience. They shift the locus of evaluation from external expert to community member — valuable for accountability to communities as well as for data quality on locally relevant indicators.

Method	What It Produces	Best For
Participatory mapping	Community-drawn maps of resources, risks, services	Baseline mapping; infrastructure access; hazard exposure
Seasonal calendar	Annual pattern of workload, income, food security, illness	Understanding livelihood seasonality; programme timing
Trend lines	Community assessment of change over time (10–20 years)	Historical context; community-perceived change on key dimensions
Venn diagram	Community's perception of local institutions and relationships	Power mapping; stakeholder analysis; governance assessment
Most Significant Change (MSC)	Stories of the most significant changes experienced	Complex programmes; capturing unexpected outcomes; human face of impact

Most Significant Change: Developed by Rick Davies (1996), MSC involves: (1) collecting stories of significant change from participants; (2) deliberate selection of the most significant stories by panels at different levels; (3) reporting back the stories selected and why. It produces rich, unexpected insights about programme impact — including negative and unintended impacts — that structured surveys miss entirely.

Caution on participation in MEL: Participatory methods can be co-opted into compliance reporting — communities "participate" in producing outputs that validate programme decisions already made. True participatory MEL requires communities to have genuine power over what gets measured, what counts as success, and what the findings mean. This is organisationally demanding and requires deliberate institutional commitment.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Digital Tools

Digital Data Collection: Tools, Trade-offs, and Indian Reality

Mobile-based data collection has transformed field MEL in India over the past decade — reducing errors, enabling real-time data review, and cutting transcription costs. But digital tools also create new constraints and risks that must be managed, particularly in low-connectivity field environments.

Tool	Type	Strengths	Limitations
KoBoToolbox	Mobile survey (free)	Offline capable; skip logic; free; widely used by NGOs	Limited dashboarding; storage limits on free tier
ODK Collect	Mobile survey (open source)	Highly customisable; open source; works offline	Requires server setup; technical capacity needed
CommCare	Case management + surveys	Longitudinal tracking; health worker workflow	Expensive; implementation complexity
DHIS2	Health information system	Government standard in many states; aggregate reporting	Not designed for programme-level MEL
Google Forms / Sheets	Survey + database	Free; familiar; quick deployment	No offline; limited validation; data security concerns

India connectivity realities: 4G connectivity in field areas — Jharkhand, Chhattisgarh, eastern UP — remains unreliable. Offline-first tools (KoBoToolbox, ODK) are essential for field teams. Syncing data when in town. Do not design data collection workflows that depend on real-time connectivity if working in Tier 3+ locations.

Data quality in digital collection: Digital tools reduce transcription error but introduce new risks: (1) Enumerators completing surveys without visiting respondents ("porch surveys"); (2) GPS data showing incorrect locations because devices don't update; (3) Metadata manipulation. GPS verification of interview location, audio recording of a sample of interviews (with informed consent), and supervisor spot-checks remain essential quality controls even with digital tools.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Administrative Data

Administrative Data: India's Underused MEL Asset

India has one of the world's richest administrative data ecosystems — NFHS, NSSO/PLFS, UDISE+, HMIS, NCRB, census, MGNREGS MIS, PM-KISAN data, and dozens of state-level systems. For many development programmes, this data can supplement or replace expensive primary data collection for baseline and context analysis.

NFHS-5

National Family Health Survey (2019–21) — district-level data on health, nutrition, women's empowerment. Free download at rchiips.org

PLFS

Periodic Labour Force Survey — quarterly and annual employment data. State-level. Essential for livelihoods baselines.

UDISE+

School education statistics by school, block, district. Enrollment, infrastructure, teacher-pupil ratios. Annual.

HMIS

Health Management Information System — facility-level health service data. Monthly data at block level publicly available.

Using administrative data in MEL: Administrative data can provide: (1) Baseline values for standard indicators before programme start; (2) Trend context — how has the indicator moved over 5–10 years before the programme?; (3) Comparison data — how does the programme area compare to similar non-programme areas?; (4) Corroboration — does primary data align with administrative data trends? Divergence signals either programme effect or data quality issues.

Limitations of administrative data: Coverage gaps in remote areas; institutional incentives to over-report (MGNREGS employment days; school attendance); significant data lags (NFHS is 3+ years old by time of release); limited disaggregation below district level for most surveys; measurement definitions that may not align with programme indicators. Use with knowledge of these limitations, not as ground truth.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Data Quality

Data Quality: The Unglamorous Foundation of Good MEL

Data quality is the most underinvested aspect of development MEL. Organisations spend significant resources collecting data, then use it without questioning its reliability or validity. Conclusions drawn from poor-quality data are worse than having no data — they create false confidence.

Five Dimensions of Data Quality (USAID DQA Framework)

Validity

Does the data actually measure what it claims to measure?

Reliability

Would the same measurement give the same result if repeated by different people?

Timeliness

Is the data current enough to inform the decision it's intended for?

Precision

Is the measurement precise enough for the decisions being made?

Integrity

Has the data been protected from manipulation or bias?

Data Quality Assurance in Practice

Enumerator training: Comprehensive training before every survey; inter-rater reliability testing; supervised practice interviews
Back-checks: Re-interview 5–10% of surveyed households by supervisor within 48 hours — compare key responses for consistency
Range checks: Built into digital survey tools — flag responses outside expected ranges (household income above 10x median; age of "child" listed as 60)
Completeness checks: Ensure no skip-pattern violations; no systematically missing fields
Data Quality Assessment (DQA): Formal review of MEL systems — USAID requires DQAs every 2 years for funded projects. Involves tracing data from field record to report to check integrity and completeness at each step.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Research Ethics

Ethics in Data Collection: Informed Consent, Privacy, and Do No Harm

MEL data collection involves collecting personal information from vulnerable people — often in situations where they have limited power to refuse and limited understanding of how their data will be used. Ethical data collection is not bureaucratic compliance — it is a fundamental obligation to the communities development programmes claim to serve.

Informed Consent

Before collecting data, respondents must: (1) understand the purpose of the study; (2) know they can refuse or stop without consequences; (3) understand how their data will be used and stored; (4) know who will have access. Consent must be genuinely informed and voluntary — not extracted through power relationships or false assurances.

Ethical Principles for Development MEL

Voluntary participation: Never imply (even indirectly) that programme benefits depend on participation in surveys. This is coercive and invalidates the consent.
Confidentiality: Individual data should not be attributable to named individuals in reports. Disaggregated small-cell data (e.g. 2 Dalit women in a group of 50) may be identifiable even without names.
Sensitive topics: Data on domestic violence, HIV status, income, and caste require special precautions — separate interviews, female enumerators for women, clear explanation of limits of confidentiality.
Data storage: Respondent-level data should be stored securely, access-restricted, and deleted when no longer needed. Under India's DPDP Act 2023, organisations have legal obligations around personal data processing.
Child data: Assent from children AND consent from parents/guardians. Extra precautions for data on child abuse, health, or identity.
Feedback to communities: Share findings with the communities who provided data. Withholding results from data subjects is an extractive practice.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 05 · Data Management

Data Management: Storing, Cleaning, and Analysing

Data collected well but stored poorly, cleaned inconsistently, or analysed superficially produces conclusions as unreliable as data collected badly. Data management is the unglamorous infrastructure of good MEL.

Data entry / import: Digital (KoBoToolbox export) or paper-to-digital. Double-entry for critical data.

Data cleaning: Range checks, logic checks, duplicate removal, missing data handling, outlier review

Variable construction: Computing indicators from raw data — score indices, disaggregations, derived variables

Analysis: Descriptive statistics; before-after comparisons; subgroup analysis; qualitative thematic coding

Storage and documentation: Raw data retained; cleaning log; codebook; version control

Data cleaning is analysis: The decisions made during cleaning — how to handle missing data, what counts as an outlier, how to code qualitative responses — are substantive analytical choices that affect conclusions. Document every cleaning decision. Share the cleaning log with evaluation users. "Raw data never lies but cleaned data always reflects assumptions."

Tools for analysis in Indian NGO contexts: Excel is the reality for most mid-size organisations — used well (pivot tables, COUNTIF, data validation) it handles most MEL analysis needs. SPSS is common in academic research collaborations. Increasingly, R and Python are used in larger NGOs with MEL specialists. The tool matters less than the analytical rigour and documentation.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Six

Evaluation Design

ImpactMojoMEL Basics 101www.impactmojo.in

Section 06 · Types of Evaluation

Types of Evaluation: Choosing the Right Question

Evaluation is not one thing — it encompasses very different questions requiring different designs and methods. Choosing the right type of evaluation starts with identifying the primary question the evaluation should answer — not with choosing a methodology.

Type	Primary Question	When
Formative	Is the programme design right? Are we reaching the right people? What needs to be adjusted?	Early implementation; piloting
Process / Implementation	Is the programme being implemented as designed? With what quality?	Mid-implementation; fidelity assessment
Outcome	Did the programme achieve its outcomes? For whom?	Near or after completion
Impact	Did the programme cause the outcomes? What would have happened without it?	After completion; requires counterfactual
Economic	Was the programme cost-effective? What is its cost per unit of outcome?	With cost data; comparison available
Developmental	Are we learning enough to adapt? Is the ToC still valid?	Complex, long-term programmes

Summative vs Formative: Formative evaluation asks "how can we improve?" — and findings are used to change the programme. Summative evaluation asks "did it work?" — and findings inform decisions about whether to continue, scale, or end. Most development evaluations try to do both at once — producing reports that are neither genuinely formative (too late to act on) nor genuinely summative (not rigorous enough to make claims about causation).

The midterm evaluation: One of the most valuable and least well-executed evaluation types. Commissioned at programme midpoint — should answer: Is implementation on track? Are early-stage outcomes appearing as expected? What course corrections are needed? Too often, midterms are mini-endlines — documenting progress rather than challenging design assumptions. The most useful midterms are adversarial — actively looking for evidence that the ToC is wrong.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 06 · RCTs

Randomised Controlled Trials: Gold Standard or Overrated?

The Randomised Controlled Trial (RCT) is the most robust design for establishing causal attribution — whether the programme caused an observed change. Random assignment to treatment and control groups makes the two groups statistically equivalent, so differences in outcomes can be attributed to the intervention. Abhijit Banerjee, Esther Duflo, and Michael Kremer won the 2019 Nobel Prize in Economics for their work applying RCTs to poverty reduction.

Counterfactual

What would have happened to programme participants if they had not received the intervention? The fundamental question in impact evaluation. The control group in an RCT provides the counterfactual — they are identical to the treatment group in expectation (due to randomisation) but did not receive the intervention.

RCTs: Strengths and Significant Limitations

Strengths

Best design for causal attribution; controls for confounders; clear counterfactual; replicable; generates sector knowledge

Limitations

Expensive; requires large sample; ethical concerns about control groups; poor external validity; answers "does it work on average" not "for whom under what conditions"; doesn't explain mechanisms

The external validity problem: An RCT shows whether an intervention worked for a specific population in a specific context. Whether that finding generalises to other populations, geographies, or scale is an entirely separate question — and one that RCTs alone cannot answer. The replication crisis in development economics (many findings that didn't replicate at scale or in different contexts) reflects this fundamental limitation. See the critique from Angus Deaton and Nancy Cartwright (2018).

ImpactMojoMEL Basics 101www.impactmojo.in

Section 06 · Quasi-Experimental

Quasi-Experimental Designs: Rigour Without Randomisation

Most development programmes cannot be randomised — it is unethical to withhold a potentially beneficial programme from a randomly assigned control group. Quasi-experimental designs use statistical or natural variation to construct a credible counterfactual without randomisation.

Design	Mechanism	Requirement
Difference-in-Differences (DiD)	Compare change in treatment vs control group over time — the "double difference"	Parallel trends assumption; baseline data for both groups
Regression Discontinuity (RDD)	Compare just above and below an eligibility threshold — treat threshold as quasi-random	Sharp eligibility cutoff; running variable measurable
Instrumental Variables (IV)	Find a variable that predicts programme participation but doesn't directly affect outcome	Valid instrument (rare); strong first stage
Propensity Score Matching (PSM)	Match treated units with untreated units with similar observable characteristics	All confounders measured; strong on observables
Synthetic Control	Construct a weighted "synthetic" comparison from multiple control units	Pre-treatment data for treatment and potential controls; few treated units

DiD in practice: Difference-in-Differences is the most commonly used quasi-experimental method in development evaluations. If a programme is rolled out in Phase 1 districts and Phase 2 districts (the Phase 2 group serving as controls during Phase 1), DiD estimates the programme effect as: (change in Phase 1 outcomes) minus (change in Phase 2 outcomes over the same period). The critical assumption — parallel trends — means the two groups would have changed similarly without the programme.

For practitioners — the key message: You don't need to implement quasi-experimental methods yourself, but you need to understand them well enough to: (1) recognise when an evaluation is claiming to show causal impact without sufficient design rigour; (2) commission evaluations that include appropriate counterfactual designs; (3) interpret findings correctly — knowing whether a result is a correlation or a causal estimate.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 06 · Non-Experimental Methods

Non-Experimental Evaluation: Theory-Based and Qualitative Approaches

Most development evaluations in India and globally are non-experimental — they cannot establish a counterfactual. They can still be rigorous, useful, and honest about what they can and cannot claim. The key is matching the evaluation design to the evaluation question — not defaulting to quantitative methods that imply causal claims they can't support.

Pre-post design (before-after): Most common non-experimental design. Measure outcomes before the programme (baseline) and after (endline). The change is attributed to the programme. The fatal flaw: cannot distinguish programme effects from secular trends, seasonal variation, or other external changes. A village's income may have risen for reasons entirely unrelated to the programme.

Process tracing: A theory-based evaluation method. Looks for observable evidence consistent with the causal mechanism claimed in the ToC — and for evidence that would rule out alternative explanations. If the ToC says "training leads to knowledge change which leads to behaviour change," process tracing looks for evidence of knowledge change in the right timeframe, then behaviour change, then tests whether knowledge change predicted behaviour change at individual level.

Contribution Analysis: An Approach for Complex Programmes

Contribution Analysis (John Mayne) is designed for complex programmes where attribution is not possible. It builds a "contribution story" — a plausible, evidence-based narrative that the programme contributed to observed change, acknowledging other factors.

State the ToC and what evidence would support or undermine it

Collect evidence on outcomes actually observed

Assess whether observed outcomes match the ToC's predictions

Identify and assess alternative explanations

Build and communicate the contribution story with confidence levels

ImpactMojoMEL Basics 101www.impactmojo.in

Section 06 · Commissioning Evaluations

Commissioning an Evaluation: What to Include in a ToR

Terms of Reference (ToR) for an evaluation are one of the most important documents in the MEL process. A well-written ToR produces relevant, useful evaluation findings. A poorly written ToR produces a report that costs significant resources and answers nobody's actual questions.

Essential Elements of a Good Evaluation ToR

Programme description and context: Theory of change, target population, geography, budget, timeline, what has been implemented so far
Evaluation purpose: Why is this evaluation being done? Who will use the findings and for what decisions?
Evaluation questions: Specific, answerable questions — not vague objectives. Maximum 4–6 questions.
Evaluation design: What approach is expected — or left to the evaluator to propose? What counterfactual design is feasible?
Data requirements: What existing data is available; what primary data is needed; any excluded methods
Deliverables and timeline: Inception report, draft, final report, presentation — with specific dates

The evaluation question must be specific: "Evaluate the programme's effectiveness" is not an evaluation question. These are: "To what extent did women in the programme increase their decision-making power over household finances compared to the baseline?" or "What implementation factors explain the difference in adoption rates between high-performing and low-performing districts?" Each is specific, answerable, and linked to a decision.

Internal vs external evaluation: Internal evaluation (by programme staff) has better access to data and context; lower cost; often more practical in recommendations. Risk: bias and lack of independence. External evaluation (by consultants) brings independence and methodological expertise; seen as more credible by donors. Risk: superficiality; context ignorance; extractive. Best practice: external evaluator with genuine partnership with internal MEL team throughout the process.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 06 · Synthesis

Evaluation Design Decision Guide

Evaluation Methods by Rigour and Feasibility for Typical Indian NGO Contexts

Source: ImpactMojo MEL guidance — illustrative positioning based on resource requirements and methodological strength

Matching Design to Question and Resources

If your primary question is...	Consider...
What happened and did we reach our targets?	Pre-post design; quantitative survey + qualitative
Did we cause this change?	RCT (if feasible) or quasi-experimental (DiD, RDD)
Why did/didn't outcomes occur?	Process evaluation; qualitative methods; contribution analysis
Is our ToC still valid?	Theory-based evaluation; outcome harvesting
What's the value for money?	Cost-effectiveness or cost-benefit analysis
How can we improve this year?	Formative evaluation; rapid qualitative assessment

Most important MEL decision: Choose your evaluation question before your method. The most common evaluation failure — using a pre-post quantitative survey regardless of the question — produces expensive data that cannot answer whether the programme caused anything or why outcomes did or didn't occur. Methodology follows question.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Seven

Attribution & Contribution

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · Attribution

The Attribution Problem: Why Causation Is Hard in Development

Attribution

The degree to which observed outcomes can be ascribed to the programme, rather than to other factors (external trends, other actors, seasonality, maturation effects). Full attribution requires knowing what would have happened in the absence of the programme — the counterfactual. In most development contexts, this is impossible to know with certainty.

The attribution gap in practice: A women's livelihoods programme reports that beneficiary household incomes rose 35% over three years. Is that the programme? A national economic upswing? Rising commodity prices? A government scheme running in the same area? The fact that incomes rose does not tell us why they rose. Without a counterfactual — data on what happened to comparable non-beneficiaries over the same period — the 35% figure is a correlation, not a programme effect.

The Attribution Spectrum

High

RCT with large sample

Random assignment eliminates selection bias. Strongest causal claim. Expensive and often infeasible.

Med+

DiD, RDD, IV

Quasi-experimental. Credible causal claim if assumptions hold. Requires strong design discipline.

Med

Matched comparison

PSM; comparison villages with similar characteristics. Weaker — only controls for observable confounders.

Low

Pre-post without comparison

Cannot rule out external factors. Shows change occurred; cannot attribute it.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · Contribution

Contribution vs Attribution: Reframing the Question

For most development programmes — especially those addressing complex, multi-causal problems — the honest question is not "did we cause this?" but "what was our contribution to this outcome, alongside other factors?" Contribution is not a lesser standard than attribution; it is a more honest one for the contexts development programmes operate in.

The question is not whether the programme made a difference but whether it made more of a difference than alternative uses of the same resources.

Paraphrasing John Mayne — contribution analysis framing

Making the Contribution Case: What Evidence Is Needed

The ToC held: Evidence that the causal pathway specified in the ToC actually played out — the intermediate steps happened in the right sequence
Other explanations addressed: Key alternative explanations investigated and ruled out or assessed for their relative importance
Beneficiary attribution: The people who changed attribute the change (at least partly) to the programme — not to something else
Dose-response: Those who received more of the intervention show more change — this is consistent with a causal story even without a control group
Timing: Change appeared after the programme started, not before — and in areas where the programme operated, not in areas where it didn't

Honest communication: A contribution claim should be stated as such: "The evidence suggests the programme was a significant contributing factor to observed income improvements, alongside favourable commodity prices and improved road access." This is more credible than claiming full attribution — and more useful for learning.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · Realist Evaluation

Realist Evaluation: What Works for Whom in What Circumstances

Realist evaluation, developed by Ray Pawson and Nick Tilley, rejects the simple "did it work?" question. Instead, it asks: "What works, for whom, in what circumstances, and why?" It is particularly useful when a programme works in some contexts and not others — and you need to understand why.

CMO Configuration

Context + Mechanism + Outcome. Realist evaluation seeks to identify which contexts activate which mechanisms to produce which outcomes. The same intervention can produce different outcomes in different contexts — because context determines whether the mechanism fires. Understanding the CMO configuration is how you build transferable knowledge about what works and when.

Realist example: A microfinance programme works well in one district (high outcome) and poorly in another (low outcome). A standard evaluation would average the two and conclude "moderate impact." Realist evaluation asks: what is different between the contexts? It finds: in the high-outcome district, women's self-help groups already existed — the programme activated a peer-accountability mechanism that was absent in the low-outcome district. The mechanism (peer accountability) was present in one context and absent in the other. This is actionable learning.

Practical value for Indian practitioners: India's diversity — across caste, language, agro-ecology, governance quality, and social norms — means context variation is extremely high. A programme that works in Tamil Nadu may fail in UP. Realist evaluation provides the analytical framework to understand and document why — and to build a knowledge base that allows programmes to be adapted to different contexts rather than failing silently.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · Systems Evaluation

Systems Change Evaluation: Measuring What Logframes Cannot

Programmes aimed at systems change — shifting policies, markets, norms, or institutional arrangements — cannot be evaluated with standard logframe-and-indicator approaches. The change they seek is not in individual outcomes but in the system that produces those outcomes. Evaluating systems change requires different questions, methods, and timelines.

What does systems change look like? Policies changed; market rules that were unjust no longer apply; norms that excluded certain groups no longer dominate; institutions that failed marginalised people now serve them. These are changes in the architecture of a system — not in individual behaviour. Measuring them requires tracing policy change, mapping power shifts, and assessing whether the system itself is more equitable or resilient.

Outcome Harvesting for systems change: Developed by Ricardo Wilson-Grau. Rather than measuring pre-specified indicators, Outcome Harvesting collects "outcomes" — changes in the behaviour, relationships, policies, or practices of actors — then works backwards to assess the programme's contribution. It captures emergent, unexpected, and systemic changes that standard evaluation misses. Used by Ford Foundation, Open Society, and many advocacy-focused programmes.

Systems Change Indicators: Examples

System	Change Type	Indicator Approach
Policy	New legislation enacted; policy implemented with budget	Policy mapping; implementation tracker; budget analysis
Market	New market actors, rules, or norms emerge	DCED standards; market system mapping at baseline and endline
Norms	Shift in what is considered acceptable behaviour	Social norms measurement tools (EMERGE); longitudinal qualitative
Power	Marginalised groups have greater voice and agency	Power analysis frameworks; Most Significant Change; network analysis
Institutional	Government systems deliver for excluded groups	Service delivery scorecards; beneficiary feedback; PEFA assessments

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · Cost-Effectiveness

Cost-Effectiveness Analysis: Getting to Value for Money

Cost-effectiveness analysis (CEA) asks: how much did it cost to achieve one unit of outcome? Combined with impact evidence, it allows comparison across programmes and informs resource allocation decisions. It is increasingly demanded by donors — but rarely done well.

CEA

Cost-Effectiveness Analysis: cost per unit of outcome. E.g. cost per child prevented from dropping out of school. Compares efficiency without monetising outcomes.

CBA

Cost-Benefit Analysis: monetises outcomes and compares to costs. Requires strong assumptions about value of outcomes. More powerful but harder to do credibly.

Calculating cost per outcome: Total programme cost (including staff time, overheads, capital costs — not just direct activity costs) ÷ number of outcome units achieved. If a programme spends ₹2 crore and 500 women complete vocational training and are employed after 1 year: cost per employed woman = ₹40,000. Is this good value? Only answerable by comparison to alternatives.

Common CEA Errors and How to Avoid Them

Incomplete cost capture: Using only direct activity costs and ignoring staff time, management overhead, and capital — produces artificially low cost-per-outcome figures
Using outputs not outcomes: "Cost per training session" is not CEA. Cost per sustained behaviour change is.
Ignoring programme quality: A cheap programme with poor quality outcomes is not more cost-effective than a more expensive one with durable outcomes
No comparison benchmark: A cost figure means nothing without a reference point — what does the same outcome cost through government schemes, or through alternative NGO approaches?

GiveWell and the accountability shift: GiveWell's rigorous CEA of global health charities has moved significant philanthropic funding towards cost-effective interventions (bed nets, vitamin A supplementation, direct cash transfers). The same pressure is now moving into Indian philanthropy — donors asking "how does your cost-per-outcome compare to the best alternative?"

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · OECD-DAC Criteria

OECD-DAC Evaluation Criteria: The Industry Standard

The OECD Development Assistance Committee (DAC) evaluation criteria are the most widely used framework for structuring development evaluation questions. Originally five criteria, a sixth — coherence — was added in 2019. Understanding them allows practitioners to structure evaluation questions and communicate findings using internationally recognised language.

Relevance

Is the programme addressing the right problem, for the right people, in the right way given the context?

Coherence

Does the programme fit with other interventions by the same organisation and by other actors? Are there synergies or contradictions?

Effectiveness

Did the programme achieve its intended outcomes and impact? To what extent?

Efficiency

Were inputs (money, time, staff) used well to achieve outcomes? Could the same results have been achieved with less?

Impact

What are the broader, longer-term effects — positive and negative, intended and unintended — on people and systems?

Sustainability

Will the benefits continue after the programme ends? Are changes institutionalised? Is the system stronger?

Using OECD-DAC in ToR: An evaluation ToR will often structure evaluation questions under each criterion: "Under Effectiveness, the evaluation will assess: to what extent did the programme achieve its outcome targets?" This provides international comparability and a complete assessment framework. But practitioners should note: not every evaluation needs to address all six — prioritise based on the decision being made.

Sustainability: the most neglected criterion: Evaluations consistently find that sustainability — whether programme gains are maintained after the programme ends — is the weakest element of most development programmes. Infrastructure built by NGOs deteriorates without maintenance systems. Behaviour changes revert when support structures are removed. Learning that was never institutionalised disappears with staff turnover. A programme that achieves its outcomes but produces nothing durable is fundamentally incomplete.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · Unintended Effects

Unintended Effects: What Programmes Don't Plan For

All programmes produce unintended effects — positive and negative. Positive unintended effects are often ignored in reporting because they weren't planned. Negative unintended effects are often actively concealed because they undermine the programme narrative. Rigorous MEL actively looks for both — because they contain the most valuable learning.

Type	Development Examples
Positive unintended	Women's SHG programme improves men's alcohol reduction (peer pressure from wives); school feeding programme improves sibling enrolment (family strategy)
Negative unintended	Microfinance increases women's access to credit but also increases domestic violence when husbands lose control of finance; WASH programme builds toilets but women still use open spaces (safety at night)
Displacement	Programme provides jobs in target area; unemployed from neighbouring areas migrate in, displacing programme beneficiaries
Dependency	Emergency food distribution over multiple years reduces households' agricultural investment — because they anticipate free food
Elite capture	Community-based targeting of benefits is captured by local elites who control access; most marginalised excluded

Do No Harm: Borrowed from medical ethics, "do no harm" in development MEL means systematically assessing whether the programme is producing negative effects on any sub-group — and designing MEL systems to detect these. Programmes that only measure their intended targets cannot see the harm they may be causing to excluded or differently-affected groups.

How to detect unintended effects: (1) Open-ended questions in endline surveys: "What, if anything, about the programme has been harmful or problematic for you or your family?"; (2) FGDs designed to surface negative experiences (requires significant trust and careful facilitation); (3) Outcome Harvesting which collects all significant changes, not just positive ones; (4) Complaints and feedback mechanisms that are actually functional and safe to use.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 07 · Synthesis

Attribution & Contribution: Key Takeaways for Practitioners

What Good Practice Looks Like

Be explicit about the level of causal claim you can make given your evaluation design — don't overstate
Use contribution framing when full attribution is not possible — it is honest, not weak
Design MEL systems to detect unintended effects — positive and negative
Consider realist questions (for whom, in what circumstances) when interpreting results
Track sustainability from the start — not as an endline afterthought
Build cost data collection into programme operations — so CEA is possible without a separate costing exercise

On overstatement: The development sector's credibility problem is substantially a product of overclaimed attribution — programmes presenting correlation as causation, before-after change as programme effect, and output counts as impact. Funders have become appropriately sceptical. Rigorous, honest causal claims — even when modest — are more credible and more useful than inflated ones.

Common Errors to Avoid

"Beneficiaries reported improved incomes" — self-report without baseline or counterfactual is not impact evidence
Presenting before-after change as the programme effect without acknowledging secular trends or external factors
Reporting the best-performing sub-group results as overall programme results
Attributing all observed change to your programme when other actors were working in the same area
Not reporting negative or null findings because they are inconvenient for donor reporting
Conducting CEA using direct activity costs only, excluding staff and overhead

The sector norm you can change: When you report findings honestly — including what didn't work, what you're uncertain about, and what may have caused harm — you signal to your funders that your positive findings are credible. The organisation that only reports success is not credible. The organisation that reports honestly is.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Eight

Learning Systems & Adaptive Management

ImpactMojoMEL Basics 101www.impactmojo.in

Section 08 · Learning Systems

The Learning Function: From Data to Decisions

Learning — using evidence from monitoring and evaluation to improve programme design and implementation — is the least developed of the three MEL functions. Most organisations have strong monitoring systems and conduct evaluations. Far fewer have systems that reliably translate evidence into changed practice.

Why learning is hard: (1) Evidence takes time to collect and analyse — often too late for the decision it was meant to inform; (2) Findings that show poor performance create institutional defensiveness, not curiosity; (3) Learning requires space for honest internal conversation — rare in high-pressure delivery environments; (4) Staff turnover means learning leaves when people leave; (5) Donors rarely fund learning activities explicitly — it's treated as overhead.

Building a Learning System: Key Components

Regular review rhythm: Monthly field review meetings; quarterly programme reviews with evidence; annual strategic reflection. Not just financial reporting.
Structured learning events: After-action reviews; pause-and-reflect sessions; peer learning across programmes
Documentation: Learning notes; lesson logs; failure reports (actively encouraged). Not just evaluation reports that nobody reads.
Decision protocols: Explicit agreements about which evidence triggers which decisions. If attendance drops below X%, we do Y. If outcome indicator is off-track at midterm, we convene a review within 30 days.
Knowledge management: Where is learning stored? Who has access? How does it inform new programme design?

ImpactMojoMEL Basics 101www.impactmojo.in

Section 08 · Adaptive Management

Adaptive Management: Adjusting Course Based on Evidence

Adaptive management is the organisational practice of systematically adjusting programme design and delivery in response to evidence from monitoring and evaluation. It is the operationalisation of the L in MEL — not a concept but a set of practices and conditions that allow an organisation to change course when evidence indicates it should.

Adaptive Management

A structured, intentional approach to learning from implementation and adjusting accordingly. It involves: (1) regular evidence review; (2) willingness to challenge assumptions; (3) authority to make changes at programme level without lengthy approval chains; (4) documentation of what was changed and why. FCDO's "Thinking and Working Politically" and USAID's "CLA" (Collaborating, Learning, Adapting) are the major frameworks.

CLA — Collaborating, Learning, Adapting (USAID): USAID's CLA framework requires implementing partners to actively collaborate with other actors (not work in silos), invest in learning (not just data collection), and deliberately adapt programmes based on evidence. CLA is assessed in project design, annual reviews, and performance management. It represents a significant shift from compliance-focused MEL to learning-focused MEL.

What prevents adaptation: (1) Donor approval requirements for any programme change — creating institutional disincentive to adapt; (2) Fear of appearing to have "failed" if original plan changes; (3) Staff incentives tied to delivery of the original workplan rather than achievement of outcomes; (4) No systematic review process — evidence exists but is never formally reviewed; (5) Leadership that doesn't create space for honest internal discussion of what isn't working.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 08 · PDIA

PDIA: Problem-Driven Iterative Adaptation

Problem-Driven Iterative Adaptation (PDIA) is an approach to development practice and evaluation developed by Matt Andrews, Lant Pritchett, and Michael Woolcock at Harvard's Building State Capability programme. It starts from a specific, locally-defined problem — not a donor-designed solution — and iterates towards solutions through rapid experimentation and learning.

Traditional Model

Plan → Implement → Evaluate

Pre-designed solution; implementation fidelity valued; evaluation at end. Assumes you know the solution in advance.

PDIA Model

Problem → Iterate → Learn → Adapt

Problem-driven; rapid cycles of action and reflection; solution emerges from local experimentation. MEL is built into each iteration.

PDIA in practice: (1) Deconstruct the problem — what are its component parts? What could be changed? (2) Push authority to the team closest to the problem — they can experiment. (3) Run rapid experiments — small, low-cost tests of potential solutions. (4) Reflect and learn — what worked? Why? (5) Iterate — don't scale until you understand what you're scaling. PDIA explicitly rejects the logframe-and-scale model for complex state capability and governance problems.

PDIA and MEL: MEL in a PDIA approach is continuous and embedded — not a periodic assessment function. Data is collected in real-time; learning loops are weekly, not annual. The team runs mini-evaluations constantly: "Did this experiment work? What does the evidence say? What do we try next?" This requires a fundamentally different MEL architecture from compliance-driven logframe monitoring.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 08 · Developmental Evaluation

Developmental Evaluation: Evaluation as a Learning Partner

Developmental Evaluation (DE), developed by Michael Quinn Patton, is designed for programmes working in complex, dynamic situations where the programme is still being developed. Unlike summative evaluation (which assesses a completed programme) or formative evaluation (which improves a defined programme), DE partners with innovators throughout the development process — using evaluation thinking to support learning and adaptation in real time.

Developmental evaluation supports innovation development to guide adaptation to emergent and dynamic realities in complex environments.

Michael Quinn Patton — Developmental Evaluation (2011)

When Developmental Evaluation Is Appropriate

Social innovation: The programme is genuinely new — no established best practice; the team is figuring out the approach as they go
Complex adaptive systems: The context is dynamic; the programme needs to change as the context changes; emergence is expected
Scaling: Moving a proven approach to a new context where it may work differently
Crisis response: Rapidly evolving situation where standard planning is impossible

DE vs standard evaluation: A standard evaluation assesses against pre-determined criteria. DE helps the programme figure out what its criteria should be as it learns. A standard evaluator stands outside the programme. A developmental evaluator is embedded within it — using systems thinking, real-time data, and facilitated reflection to support the team's learning. Not all funders accept DE instead of logframe-based MEL — negotiate early.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 08 · Learning Loops

Learning Loops: Single, Double, and Triple

Chris Argyris's theory of organisational learning distinguishes between types of learning by how deeply they challenge existing assumptions. Development MEL systems typically enable single-loop learning at best. Moving to double and triple loop learning is what distinguishes genuinely learning organisations from compliance-oriented ones.

Single-Loop Learning

Detect error; correct within existing norms and strategy. "Our coverage is low — let's hire more field staff." Does not question whether the approach is right.

Double-Loop Learning

Detect error; question the norms and strategy themselves. "Our coverage is low — is field-based delivery the right model? Are we reaching the right people?" Challenges underlying assumptions.

Triple-Loop Learning

Question how we learn — our learning systems themselves. "Why do we consistently fail to detect coverage problems early? What is wrong with our MEL system?" Organisational epistemology.

Why development organisations stay at single-loop: Double and triple-loop learning require challenging leadership decisions, funding model assumptions, and organisational identity. "Our approach to community mobilisation isn't working" implies that senior leadership chose the wrong strategy — which creates political risk. MEL that only produces single-loop learning is structurally safe and structurally limited.

Creating conditions for double-loop learning: (1) Leadership that explicitly invites challenge: "What are we getting wrong? What should we do differently?"; (2) Psychological safety — staff must be able to report problems without career consequences; (3) External facilitation of strategic reviews — an external perspective breaks groupthink; (4) Regular assumption testing built into MEL plan; (5) Failure reports rewarded, not punished.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 08 · Knowledge Management

Knowledge Management: Making Learning Institutional

Knowledge management (KM) is the set of systems and practices through which organisations capture, organise, share, and use the knowledge they generate. In development organisations, learning that is not institutionalised leaves when people leave. High staff turnover — common in the sector — means the same mistakes are made repeatedly, the same questions answered from scratch, and the same communities re-surveyed for information already collected.

40–60%

Annual staff turnover rates in Indian development NGOs in some studies — meaning learning leaves faster than it is institutionalised

Exit gap

Most organisations have no structured process for capturing knowledge when experienced staff leave — it walks out the door

Practical KM for Development Organisations

Programme documentation: Not just reports — decision logs, assumption tests, implementation lessons, what changed and why. Stored accessibly, not in email chains.
After-action reviews: Structured reflection at end of each major activity. What went well? What didn't? What do we do differently next time? 60 minutes; documented.
Learning notes: Short (1–2 page) structured summaries of key lessons from programmes — not long evaluation reports. Shareable across teams.
Onboarding knowledge transfer: New staff given structured access to programme history, decisions made, lessons learned — not starting from scratch.
Exit interviews: Structured departure process to capture what leaving staff know that isn't written down.
Cross-programme learning: Platforms for sharing what works across programmes within an organisation — not siloed learning.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 08 · Feedback Mechanisms

Community Feedback Mechanisms: Closing the Accountability Loop

Community feedback mechanisms — systems through which programme participants can communicate their experiences, concerns, and suggestions to the organisation — are a critical but widely underdeveloped component of learning systems. They represent accountability downward (to communities) rather than upward (to donors).

ALNAP community feedback research: The Active Learning Network for Accountability and Performance in Humanitarian Action has documented systematically that beneficiary feedback mechanisms are: (1) rarely designed with communities; (2) often not accessible to the most marginalised; (3) frequently not responded to — making them performative rather than functional; (4) almost never used to change programme design. Feedback without response is not accountability.

Designing Functional Feedback Mechanisms

Accessible: Multiple channels (hotline, SMS, suggestion box, community interface person, regular feedback meetings) for different literacy levels and comfort with technology
Confidential: Mechanisms that don't expose complainants to retaliation — anonymous options essential for sensitive feedback
Responsive: Every complaint/feedback acknowledged; substantive response provided; feedback loop closed. If you don't respond, the mechanism will stop being used.
Disaggregated: Track who is using the mechanism — if it only reaches literate, male, higher-caste community members, it's not reaching the most marginalised
Integrated with MEL: Feedback data reviewed at programme review meetings; patterns analysed; decision trails documented

Core Humanitarian Standard: For humanitarian and many development organisations, the Core Humanitarian Standard (CHS) — a set of nine commitments to quality and accountability — includes community participation and feedback as a central requirement. CHS verification is increasingly required by major donors.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Nine

Communicating Evidence

ImpactMojoMEL Basics 101www.impactmojo.in

Section 09 · Reporting

Writing MEL Reports That Are Actually Used

Most development evaluation reports are not read beyond the executive summary — and often not even that. They are produced for compliance, filed, and forgotten. The discipline of writing reports that generate genuine engagement and action is separate from the discipline of generating good data — and equally important.

Why reports don't get used: (1) Too long — 80-page reports are never read in full; (2) Written for evaluators, not for decision-makers — technical language, excessive methodology description, insufficient actionable findings; (3) Finding buried in passive voice and hedging; (4) Recommendations are vague ("further strengthen community engagement") not actionable ("replace monthly community meetings with fortnightly meetings focused on X"); (5) No follow-up on whether recommendations were implemented.

The key question before writing: Who will use this report? For what decision? That question should determine the structure, language, length, and findings emphasis — not the template provided by the donor or evaluator's preferences.

MEL Report Writing: Key Principles

Lead with findings, not methodology: The executive summary should state findings first; methods in an appendix for those who need them
Be direct about significance: Not "data suggests a possible trend toward improved outcomes" but "women's agricultural income increased 28% compared to a 6% increase in the comparison group"
Show uncertainty honestly: Confidence levels; sample size limitations; alternative interpretations — but state them clearly, not as an escape hatch
Make recommendations specific and assigned: "Programme team should revise the training curriculum by Q2 2025 to address low retention rates in Module 3" — not "training quality should be improved"
Balance positive and negative: A report with no negative findings is not credible. A report with useful negative findings, contextualized and with recommendations, is valuable.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 09 · Data Visualisation

Data Visualisation for MEL: Making Evidence Legible

Well-designed data visualisation makes evidence accessible to audiences who won't read tables — programme staff, community representatives, government officials, and media. Poorly designed visualisation obscures findings, misleads, or simply confuses. Data visualisation is a communication skill, not a technical one.

Chart Type	Best For	Common Misuse
Bar chart	Comparing quantities across categories	Too many bars; non-zero axis baseline inflating differences
Line chart	Trends over time; continuous data	Used for unconnected categories; cherry-picked time period
Pie chart	Parts of a whole — with ≤5 segments	Too many segments; 3D effects that distort area
Scatter plot	Relationship between two variables	Without a regression line or correlation statistic when one is implied
Map / choropeth	Geographic variation in indicators	Using raw counts instead of rates; misleading colour scales

The non-zero baseline error: The most common data visualisation manipulation in development reporting — starting a bar chart's Y-axis at a value above zero. A change from 40% to 44% shown on an axis from 38% to 46% looks like a dramatic improvement. On a proper 0–100% axis, it is modest. Always check that progress visualisations use zero-origin axes unless the scale is genuinely irrelevant to interpretation.

Show disaggregation visually: Disaggregated data should be disaggregated visually — not averaged out. If women's outcomes and men's outcomes differ significantly, show them in separate bars or separate lines. Averaging hides systematic exclusion. Side-by-side comparisons of subgroup data are one of the most powerful ways to make equity issues legible to non-technical audiences.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 09 · Audiences

Communicating Evidence to Different Audiences

Audience	What They Care About	Format & Length	Language & Framing
Donor / funder	Compliance with indicators; value for money; lessons for future funding	Formal report, 30–60 pages; PPT summary; structured findings	Technical; aligned to donor frameworks (OECD-DAC); quantitative where possible
Programme leadership	What to do differently; strategic decisions; risks to address	Management brief, 5–10 pages; presentation; actionable recommendations	Direct; decision-oriented; findings ranked by significance
Field teams	What is working and what isn't in their area; practical changes	Visual one-pager; verbal debrief; local-language summary	Simple; concrete examples from their context; focus on what they control
Communities	What happened with their data; whether they will be heard; practical changes to the programme	Community meeting; visual display; verbal presentation in local language	Non-technical; participatory; feedback invited; specific to their context
Government	Alignment with government programmes; evidence for policy; scalability	Policy brief, 2–4 pages; meeting presentation	Aligned to government priorities; evidence quality emphasised; scale potential
Media / public	Human interest; sector significance; accountability	Press release; infographic; story-led summary	Plain language; specific numbers; compelling case studies

ImpactMojoMEL Basics 101www.impactmojo.in

Section 09 · Policy Influence

Using MEL Evidence for Policy Influence

Good MEL evidence can influence government policy, sector practice, and funding priorities — but only if it is communicated to the right people at the right time in the right form. Evidence does not speak for itself. Policy influence requires understanding how policy is actually made — which is rarely through rational evidence uptake alone.

The evidence-policy gap: Policy is shaped by politics, institutional interests, existing commitments, and power — not just evidence. Research by Overseas Development Institute shows that "what works" evidence is rarely the primary driver of policy decisions, even when available. Evidence enters policy through windows of opportunity, champions within government, and fit with the political agenda. MEL practitioners who understand this use different strategies than those who simply publish findings.

Evidence-to-Policy: What Actually Works

Build relationships early: Engage government officials during programme design, not after evaluation. Co-creation produces champions.
Frame around government priorities: Evidence that shows how your programme advances the government's stated goals gets further than evidence that challenges them
Produce policy-ready formats: Two-page policy briefs; structured presentations; comparison with government programme data — not 80-page evaluation reports
Identify policy windows: Budget cycles; new government programmes being designed; elections and new administration priorities — these are when evidence gets heard
Build coalitions: Evidence championed by multiple organisations, sector networks, and respected institutions is harder to ignore than a single NGO's self-evaluation

ImpactMojoMEL Basics 101www.impactmojo.in

Section Ten

MEL in Practice: Common Failures

ImpactMojoMEL Basics 101www.impactmojo.in

Section 10 · Common Failures

10 Common MEL Failures — Numbers 1 to 5

1. MEL designed after programme, not before

Retrofit MEL produces indicators disconnected from the programme's actual theory and no baseline data. MEL must be co-designed with the programme at inception.

2. No baseline data collected

Without a baseline, the endline shows conditions but not change. Reconstructing a baseline post-hoc from recall or administrative data is methodologically weak.

3. Indicator overload

50+ indicators produce data that cannot be collected reliably, analysed meaningfully, or used for decisions. The discipline of selecting fewer, better indicators is the harder MEL skill.

4. Outputs reported as outcomes

Training counts, meeting numbers, and beneficiary reached figures appear in outcome rows. Donors and organisations both collude in this — it's easier to count than to measure change.

5. Data collected but not used

Field teams collect data; it goes to headquarters; it appears in a report; nobody makes a decision based on it. The decision to collect data must be preceded by a decision about how it will be used.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 10 · Common Failures

10 Common MEL Failures — Numbers 6 to 10

6. Evaluation findings ignored

The evaluation is conducted; the report is submitted; nothing changes. No action plan; no follow-up; no accountability for recommendations. The most expensive MEL failure because significant resources were spent for no organisational benefit.

7. Lack of disaggregation

Aggregate data hides differential outcomes. A programme may show strong average gains while systematically excluding Dalit women, conflict-affected households, or the poorest quintile. Disaggregation by gender, caste, disability, and geography is not optional.

8. No learning from failure

Programmes that don't work are terminated or quietly redesigned without documented lessons. The sector lacks a culture of structured failure learning — which means the same failures recur across organisations and contexts.

9. MEL seen as the MEL officer's job

When MEL is siloed in one person or department, programme staff do not own the data, do not question findings, and do not integrate learning into their work. MEL as an organisational function requires whole-team ownership.

10. Confusing compliance MEL with learning MEL

The MEL system is designed entirely around donor reporting requirements. It produces data in the formats funders need but generates no internally useful learning. The system treats MEL as a cost of doing business, not as a driver of programme quality.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 10 · Organisational Culture

MEL Culture: When the Problem Is Not Technical

Many MEL failures are not caused by poor technical skills or inadequate tools. They are caused by organisational culture — the unwritten rules about what is safe to say, what gets rewarded, and what is allowed to fail. Technical MEL skills are easier to address than cultural ones — but cultural change is where sustainable MEL improvement lies.

Culture eats strategy for breakfast. And it eats MEL systems for lunch.

ImpactMojo adaptation — based on Drucker's observation on strategy and culture

Signs of a Healthy MEL Culture

Leadership publicly acknowledges what isn't working and asks what the evidence suggests they do differently
Staff raise problems with current approaches in review meetings without fear of consequence
Negative evaluation findings are presented to funders with analysis and proposed responses — not buried in appendices
Programme changes are documented with the evidence that prompted them
MEL is discussed at all-staff meetings, not only in MEL team meetings

Signs of a Compliance MEL Culture

Reports are written to satisfy donors, not to generate internal learning
Field staff see data collection as additional burden, not as part of their work
Evaluation findings are reviewed once at a meeting and then filed
Staff are afraid to report that targets are not being met

ImpactMojoMEL Basics 101www.impactmojo.in

Section Eleven

MEL for Different Contexts

ImpactMojoMEL Basics 101www.impactmojo.in

Section 11 · Humanitarian MEL

MEL in Humanitarian Contexts: Speed, Uncertainty, and Accountability

Humanitarian MEL operates under conditions that make standard development MEL difficult or impossible: rapid onset crises, displaced populations, insecurity, extremely short timelines, and high stakes for both beneficiaries and organisations. Methods must be adapted to this context — not simply transplanted from stable development settings.

Key differences from development MEL: (1) No baseline is often possible — populations were displaced before data collection; (2) Population movement makes tracking individuals impossible; (3) Standard survey methods are inappropriate in insecure or chaotic contexts; (4) Real-time accountability to beneficiaries matters most when they are most vulnerable; (5) Evaluation must happen faster to be useful for ongoing response decisions.

Adapted MEL Methods for Humanitarian Response

Rapid needs assessment: Fast, often phone-based surveys with limited sampling rigour — used for operational decisions, not impact claims
MEAL (Monitoring, Evaluation, Accountability, Learning): The addition of "Accountability" explicitly recognises feedback mechanisms as a core function, not an add-on
Real-time monitoring: Weekly or daily output tracking; early warning indicators; operational dashboards
Community consultation: FGDs and KIIs conducted rapidly; findings fed back into response design within days, not months
Sphere standards: Minimum standards for humanitarian response (water, food, shelter, health) provide benchmarks for rapid assessment without full baseline data
Inter-agency coordination: In large-scale responses, MEL is coordinated through clusters (UN coordination mechanism) to avoid duplication and ensure compatible data

ImpactMojoMEL Basics 101www.impactmojo.in

Section 11 · Advocacy MEL

MEL for Advocacy and Policy Change

Advocacy programmes — those seeking to change policies, laws, norms, or resource allocations — are among the hardest to evaluate. The change they seek is non-linear, long-term, and shaped by political dynamics outside the organisation's control. Standard logframe indicators ("number of policy briefs produced") cannot capture whether advocacy is succeeding.

Why advocacy MEL is different: Attribution is nearly impossible — policy change results from multiple actors, political windows, and historical context. Causation is deeply contested. Timelines are long (policy change may take 10–15 years). The "beneficiaries" are often future populations, not current programme participants. And the most important changes are often relationships, influence, and agenda-setting — all hard to measure.

Advocacy MEL Approaches

Milestones tracking: Key policy milestones tracked over time — bill introduced; committee hearing; ministerial meeting; government consultation launched; policy adopted; budget allocated
Influence mapping: Who are the key decision-makers? What is the organisation's relationship with each? Has it changed? Periodic stakeholder relationship mapping.
Policy monitoring: Track whether target policies are implemented as enacted — the gap between policy adoption and implementation is often where advocacy work falls short
Most Significant Change: Captures unexpected advocacy wins and losses; useful for understanding what combination of factors opened or closed policy windows
Contribution narrative: Annual assessment of the organisation's contribution to observed policy changes — acknowledging other actors and factors

ImpactMojoMEL Basics 101www.impactmojo.in

Section 11 · Small NGO MEL

MEL for Small NGOs: Proportionate, Practical, and Useful

Small organisations — working on budgets under ₹2 crore, with no dedicated MEL staff — need MEL systems that are proportionate to their scale and capacity. The same principles apply, but the tools and approaches must be simpler, cheaper, and more embedded in normal programme operations.

Minimum viable MEL for a small NGO: (1) A clear, one-page Theory of Change; (2) 5–8 indicators with simple, defined data sources; (3) A baseline — even a light-touch one using existing secondary data and a small FGD; (4) Quarterly data review built into team meetings; (5) Annual reflection on whether the approach is working; (6) One learning note per year capturing what changed and why. This is MEL. It doesn't require a logframe of 50 indicators or a dedicated MEL officer.

Free MEL Resources for Small Organisations

Resource	What It Provides	Access
BetterEvaluation.org	Comprehensive, free database of evaluation methods, tools, and examples	betterevaluation.org — free
3ie Evidence Portal	Systematic reviews and impact evaluations; sector evidence base	3ieimpact.org — free
KoBoToolbox	Free mobile data collection; up to 10,000 submissions/month	kobotoolbox.org — free for NGOs
NFHS district data	Baseline data for health and women's empowerment indicators at district level	rchiips.org — free download
Harvard PDIA toolkit	PDIA approach materials; guides for problem-driven adaptive work	buildingstatecapability.hks.harvard.edu — free

ImpactMojoMEL Basics 101www.impactmojo.in

Section 11 · Government MEL

MEL for Government Programmes: Scale, Politics, and MIS

Government programmes in India operate at a scale where MEL faces unique challenges. A scheme reaching 100 million beneficiaries cannot be evaluated with a single pre-post survey. Political incentives create systematic pressure to report success. And the line between MEL as accountability and MEL as political performance is often blurred.

India's government MEL landscape: NITI Aayog's DMEO evaluates flagship central schemes — with varying quality and independence. The Planning Commission's evaluation function (now DMEO) has historically been more candid than ministry self-evaluations. State governments have development monitoring units of varying capacity. DARPG runs the Centralized Public Grievance Redress and Monitoring System (CPGRAMS) — a national citizen feedback mechanism that, in principle, provides real-time demand-side data on service delivery.

Working with Government MIS for MEL

MGNREGS MIS: Real-time data on employment, wages, works. Available at gram panchayat level at nrega.nic.in. Incentives for over-reporting known; useful for trend analysis.
PM-KISAN dashboard: Beneficiary-level data on cash transfer receipt. Can be used to verify targeting against other data sources.
HMIS: Monthly facility-level health service data. Block and district level available. Significant reporting gaps in remote areas.
UDISE+: Annual school-level education data. Enrollment, retention, infrastructure, teacher data. High coverage; some data quality issues.
Aspirational Districts: Delta ranking system comparing district performance on 49 indicators across 5 sectors — publicly available, updated monthly. Useful for NGOs working in aspirational districts to contextualise their work.

ImpactMojoMEL Basics 101www.impactmojo.in

Section Twelve

Further Reading & Resources

ImpactMojoMEL Basics 101www.impactmojo.in

Section 12 · Core Texts

Foundational Texts in MEL

Theory of Change & Planning

Weiss, C.H. (1995) — "Nothing as Practical as Good Theory" — foundational paper on theory-based evaluation
Mayne, J. (2001) — Contribution Analysis: An Approach to Exploring Cause and Effect — Treasury Board Canada
Pawson, R. & Tilley, N. (1997) — Realistic Evaluation — Sage Publications
Patton, M.Q. (2011) — Developmental Evaluation — Guilford Press

Impact Evaluation

Gertler et al. (2016) — Impact Evaluation in Practice (2nd ed.) — World Bank; free PDF download
Banerjee & Duflo (2011) — Poor Economics — random experiments and poverty; accessible for non-economists
Deaton & Cartwright (2018) — "Understanding and Misunderstanding Randomized Controlled Trials" — Social Science & Medicine; the key RCT critique
Andrews, Pritchett & Woolcock (2017) — Building State Capability — PDIA framework; free from Harvard

India-Specific MEL Resources

DMEO evaluations — niti.gov.in/dmeo — evaluations of flagship schemes; publicly available; variable quality but valuable for understanding government assessment of its own programmes
3ie India portfolio — 3ieimpact.org — systematic reviews and impact evaluations with India focus; particularly strong on health, WASH, and agriculture
IDInsight India work — idinsight.org — applied impact evaluation and MEL advisory with several India NGO case studies
IFMR LEAD — ifmrlead.org — research on financial inclusion, microfinance, and rural livelihoods; strong quantitative evaluations in Indian contexts

Open Access Journals & Platforms

Journal of Development Effectiveness — peer-reviewed; broad MEL methods and findings
BetterEvaluation.org — the most comprehensive free MEL methods database
ALNAP — alnap.org — humanitarian MEL; accountability research

ImpactMojoMEL Basics 101www.impactmojo.in

Section 12 · Key Concepts

MEL Vocabulary: 20 Terms Every Practitioner Should Know

Term	Definition
Theory of Change	Causal hypothesis about how programme activities lead to intended impact, with explicit assumptions
Logframe	4×4 planning matrix linking hierarchy of results to indicators, means of verification, and assumptions
Counterfactual	What would have happened in the absence of the programme — the fundamental question in impact evaluation
Attribution	Degree to which outcomes can be ascribed to the programme rather than to other factors
Contribution	The programme's role in producing outcomes alongside other factors — honest alternative to attribution claims
Baseline	Pre-programme measurement of outcome indicators, collected before the intervention begins
SMART indicator	Specific, Measurable, Achievable, Relevant, Time-bound
Goodhart's Law	When a measure becomes a target, it ceases to be a good measure
RCT	Randomised Controlled Trial — gold standard experimental design for causal attribution
DiD	Difference-in-Differences — quasi-experimental design comparing change in treatment vs control groups over time

Term	Definition
OECD-DAC criteria	Six evaluation criteria: relevance, coherence, effectiveness, efficiency, impact, sustainability
Outcome harvesting	Evaluation method that identifies outcomes that occurred and works backwards to assess contribution
MSC	Most Significant Change — participatory method collecting stories of significant change from participants
Realist evaluation	Context + Mechanism + Outcome — asks what works for whom in what circumstances
PDIA	Problem-Driven Iterative Adaptation — iterative, learning-based approach to complex development challenges
CEA	Cost-Effectiveness Analysis — cost per unit of outcome achieved
Data quality	Five dimensions: validity, reliability, timeliness, precision, integrity (USAID DQA framework)
Formative evaluation	Evaluation during implementation to improve the programme — contrasted with summative (did it work?)
Adaptive management	Structured practice of adjusting programme design based on evidence from monitoring and evaluation
MEAL	Monitoring, Evaluation, Accountability, and Learning — adds explicit accountability to communities

ImpactMojoMEL Basics 101www.impactmojo.in

Section 12 · Tools Reference

MEL Tools Reference: What to Use When

MEL Stage	What You Need	Free Tools / Resources	When to Use It
Design	ToC development; results framework; indicator selection	USAID Iris indicator library; BetterEvaluation; ODI's RAPID framework	Before programme start — non-negotiable
Baseline	Survey design; sampling; data collection; analysis	KoBoToolbox; ODK; NFHS district data; Census; PLFS	Before programme activities begin with target population
Monitoring	Regular output tracking; MIS; field data collection	KoBoToolbox; Google Sheets; DHIS2; CommCare	Ongoing throughout programme — monthly at minimum
Qualitative assessment	FGD guides; KII guides; MSC collection; observation protocols	BetterEvaluation method database; ODI RAPID qualitative tools	Quarterly or as needed for deeper understanding
Mid-term evaluation	Mixed methods; quantitative progress review; qualitative deep dive	OECD-DAC criteria; process tracing templates; contribution analysis guide	At programme midpoint — typically Year 2 of a 4-year programme
Endline evaluation	Full outcome survey; comparison analysis; economic analysis	Impact Evaluation in Practice (World Bank — free); GiveWell CEA framework	At or near programme close
Learning	After-action review templates; learning notes; KM platform	USAID CLA toolkit; Harvard PDIA materials; ALNAP learning tools	Ongoing; structured review events quarterly and annually

ImpactMojoMEL Basics 101www.impactmojo.in

Section 12 · India Data

India's Open Data Ecosystem for MEL Practitioners

India has an unusually rich open data ecosystem for a middle-income country. Development practitioners who know where to look can access district-level baseline data, programme MIS data, and administrative statistics that dramatically reduce the cost of evidence-informed MEL — without primary data collection for every indicator.

Key National Data Sources

Source	Coverage	Access
data.gov.in	National open data portal — thousands of government datasets across sectors	data.gov.in — free
NSO / MoSPI	National Statistics Office — NSSO surveys, census, PLFS, CES	mospi.gov.in — free microdata
IIPS / NFHS	National Family Health Survey — health, nutrition, women's empowerment by district	rchiips.org — free download
Devdatalab	Economic research data for India — Asher, Novosad and colleagues	devdatalab.org — free
SHRUG	Socioeconomic High-resolution Rural-Urban Geographic dataset for India — village-level data 1990–2011	devdatalab.org/shrug — free

Sector-Specific Dashboards

nrega.nic.in — MGNREGS MIS; panchayat-level employment and wage data; real-time
pmkisan.gov.in — PM-KISAN beneficiary data; state-wise dashboard
udiseplus.gov.in — School-level education data; enrollment, retention, infrastructure
hmis.nhp.gov.in — Monthly health facility data; block and district level
bharatkosh.gov.in — Government receipts data useful for budget tracking
pib.gov.in — Press Information Bureau — official government programme announcements and data releases

Using geospatial data: BHUVAN (bhuvan.nrsc.gov.in) provides satellite imagery and GIS data for India. Combined with socioeconomic indicators, it allows mapping of programme coverage against poverty, infrastructure, and climate risk — increasingly used in targeting and evaluation.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 12 · MEL Careers

MEL as a Career: Roles, Skills, and Growth Pathways

MEL has grown from a compliance function into a recognised professional domain within the development sector. Dedicated MEL roles exist at every level — from field MEL officer to MEL director to research and evaluation specialist. The skills required are genuinely cross-disciplinary: research methods, data analysis, programme design, communication, and facilitation.

Role Level	Typical Responsibilities	Skills Needed
Field MEL Officer	Data collection; tool administration; community interaction; data entry and basic analysis	Survey administration; KoBoToolbox/ODK; local language; data entry
MEL Coordinator / Analyst	Indicator tracking; report drafting; results framework management; basic statistical analysis	Excel/SPSS; report writing; logframe management; qualitative methods
MEL Manager	MEL system design; evaluation commissioning; donor reporting; team management	Evaluation design; quantitative and qualitative methods; ToC facilitation; donor frameworks
Research / Evaluation Specialist	Impact evaluation design; advanced statistical analysis; methods development; sector knowledge generation	R/Stata/Python; experimental and quasi-experimental methods; academic writing; research ethics

Skills gap in Indian MEL: The sector faces a shortage of practitioners who combine: (1) strong quantitative skills (sample size calculation, statistical analysis, quasi-experimental methods); (2) qualitative research skills (FGD facilitation, thematic analysis); (3) programmatic understanding (knowing what MEL findings mean for implementation); and (4) communication skills (translating findings for non-technical audiences). Most practitioners are strong on 1 or 2 of these — few on all four.

Building your MEL skills: ImpactMojo's MEL flagship course goes significantly deeper than this 101 deck — covering advanced evaluation methods, statistical analysis for MEL, and hands-on logframe and ToC construction. DevDiscourses has 500+ open-access papers including the core MEL literature cited in this deck. All free at impactmojo.in.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 12 · Emerging Trends

Where MEL Is Heading: Emerging Trends for Practitioners

AI and Machine Learning in MEL

Satellite data analysis, natural language processing for qualitative data coding, predictive analytics for programme targeting. Already used by IDInsight and J-PAL for rapid assessment. Raises new data ethics questions around consent and algorithmic bias.

Decolonising Evaluation

Growing movement to shift evaluation design, questions, and interpretation from Northern donor frameworks to locally-led standards. AEA equity statement; African Evaluation Association; Centre for Development Innovation's work on locally-led evaluation. Relevant for Indian organisations asserting their own evaluation standards against FCDO and USAID requirements.

Real-Time Feedback Systems

SMS, WhatsApp, and IVR-based rapid feedback mechanisms. Collect beneficiary experience data continuously rather than in annual surveys. Ushahidi, Feedback Labs, and several India-specific platforms. Raises response capacity challenges — feedback collected but not responded to is worse than none.

Equity-Focused MEL

Moving beyond gender disaggregation to full intersectional analysis — caste, disability, migration status, age. Emerging tools for measuring exclusion (Social Inclusion Index), disability-inclusive MEL frameworks from CBM, and caste-disaggregated data collection in India-specific contexts.

Open Data and Transparency

Growing norm of publishing evaluation findings and data — not just positive reports. AidData, DevTracker, and India-specific portals increase transparency. IATI standard for aid data transparency increasingly required by international donors. Growing evidence that organisations that publish negative findings build more credibility.

Climate-Adaptive MEL

Development outcomes are increasingly disrupted by climate events — droughts, floods, heatwaves. MEL systems need climate risk integration: tracking how climate events affect programme outcomes; building climate vulnerability into baseline data; distinguishing programme effects from climate-driven changes.

ImpactMojoMEL Basics 101www.impactmojo.in

Section 12 · Practitioner Checklist

MEL Practitioner Self-Assessment: Where Are You?

Foundation Level

Can explain MEL as an integrated system, not three separate functions
Can distinguish outputs from outcomes from impact — with examples
Can explain why a Theory of Change matters and what a good one includes
Can apply SMART criteria to assess or improve an indicator
Can name at least 3 types of evaluation and when each is appropriate
Can explain attribution and why claiming it is usually unjustified in development contexts

Intermediate Level

Can design a minimum viable MEL plan for a programme with 2–3 outcomes
Can explain DiD and RCT — their logic, assumptions, and limitations
Can design gender-responsive indicators that go beyond sex disaggregation
Can write evaluation questions that are specific and answerable

Advanced Level

Can commission a rigorous external evaluation and manage the evaluator relationship
Can design a quasi-experimental evaluation for a programme with phased rollout
Can facilitate a participatory ToC process with programme staff and community members
Can analyse and present evaluation findings to different audiences — technical and non-technical
Can assess and improve an organisation's MEL culture, not just its MEL tools

Next steps: ImpactMojo's MEL flagship course covers the intermediate and advanced levels in depth — with case studies, MCQs, and practical tools. Free at impactmojo.in/courses/mel/

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · MEL Quick Diagnostic

Quick MEL Diagnostic: Assessing Any Programme's MEL System

Use these ten diagnostic questions to rapidly assess the quality of a MEL system — whether you are designing a new programme, reviewing an existing one, or evaluating a partner organisation's MEL capacity.

#	Diagnostic Question	Red Flag if...	Good Sign if...
1	Does the programme have an explicit Theory of Change?	It's a diagram with no explanatory text, or doesn't exist	It's written, was developed with field staff, and is used in reviews
2	How many indicators are in the results framework?	More than 30 indicators total	Under 20, with clear rationale for each
3	Was baseline data collected before the programme?	No baseline; or collected 6+ months after start	Collected before first beneficiary contact; disaggregated
4	Are outcomes measured, or only outputs?	All indicators are training counts, meetings held, materials distributed	At least 3 outcome-level indicators with measurement plan
5	Who uses the monitoring data?	"We send it to the donor in the quarterly report"	"We review it monthly in team meetings and it triggers decisions"
6	Has an external evaluation been conducted?	Never; or the "evaluation" was a self-assessment with no independence	External evaluation with ToR; findings shared with management; action plan exists
7	Are negative findings shared with stakeholders?	Reports show only achievements against targets	Reports include what didn't work and why
8	Is data disaggregated by gender and other equity dimensions?	All data is aggregate; no breakdown by gender, caste, geography	Standard disaggregation by gender; additional by caste/disability where relevant
9	Is there a functional community feedback mechanism?	A suggestion box that was installed but nobody monitors	Multiple channels; response protocols; data reviewed in programme meetings
10	When the programme last changed because of evidence?	"I can't think of a time" or "the donor wouldn't allow it"	Specific example in the last 12 months, with documentation

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Sample Logframe

Sample Logframe: Women's Livelihoods Programme

Level	Statement	Indicators	Means of Verification	Assumptions
Goal / Impact	Improved economic security and wellbeing for women in target districts of MP	% women HHs below poverty line (NSSO definition); Women's MPI score	Household survey (endline Y5); MPI calculation from survey	Broader economic environment remains stable; no major shocks
Purpose / Outcome	Women have sustained income from own enterprises, with independent control over earnings	% women with own enterprise operational at 18 months; % women with independent bank account; Avg monthly net income (₹)	Enterprise tracking survey (Y2, Y4); bank account verification; income survey	Markets remain accessible; women retain control over earnings; husbands supportive
Output 1	Women trained in business skills and linked to credit	No. women completing 6-week training; % completing with 70%+ attendance; % accessing credit within 3 months	Training attendance register; assessment records; MFI disbursement records	Women can attend training (mobility; childcare); MFI partnership functional
Output 2	Market linkages established for trained women	No. buyer linkages established; % women with at least 1 formal buyer relationship	Sales records; buyer contracts; field observation	Buyers maintain commitment; transport infrastructure accessible
Activities	Training delivery; business plan support; market linkage facilitation; ongoing mentoring	Inputs: ₹1.8Cr budget; 6 field staff; 3 district offices	Financial reports; HR records	Adequate staff recruited and retained; budget released on schedule

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Sample MEL Plan

Sample MEL Plan Structure: Women's Livelihoods Programme

Indicator	Level	Definition	Baseline	Y2 Target	Y4 Target	Data Source	Frequency	Responsible
Women HHs below poverty line (%)	Impact	HHs where women are primary earner, below NSSO poverty line	48%	—	35%	HH survey (primary)	Baseline; Y4	MEL Manager
Women with own enterprise operational at 18mo (%)	Outcome	Enterprise operating with positive revenue for 3+ consecutive months at 18-month mark	12%	55%	65%	Enterprise tracker (primary)	Quarterly	MEL Coordinator
Women with independent bank account (%)	Outcome	Account in own name with independent access, verified by passbook review	31%	70%	80%	Passbook verification	Y1; Y2; Y4	Field MEL Officer
No. women completing training	Output	Completed 6-week course with ≥70% attendance, as per attendance register	0	600	1,200	Training register	Monthly	Field Officer
% women with credit access within 3mo	Output	Loan disbursed by partner MFI within 3 months of training completion	0%	60%	70%	MFI records	Quarterly	MEL Coordinator

The full MEL plan also includes: data collection tools for each indicator; quality assurance protocols; reporting schedule; budget for MEL activities (allocated at 7% of programme budget); and a learning calendar with dates for quarterly review meetings, mid-term evaluation (Month 24), and final evaluation (Month 46). MEL plan reviewed annually against implementation realities.

ImpactMojoMEL Basics 101www.impactmojo.in

ImpactMojo · Related Courses

Go Deeper: ImpactMojo Courses for MEL Practitioners

Flagship Course

Monitoring, Evaluation & Learning

The comprehensive MEL flagship course — covering advanced evaluation design, indicator development, quantitative and qualitative methods, and learning systems. Multiple modules, MCQs, case studies, and lexicon.

impactmojo.in/courses/mel/ →

Flagship Course

Data Visualisation for Development

How to present MEL evidence visually — chart selection, data storytelling, presenting to different audiences, and common visualisation errors in development reporting.

impactmojo.in/courses/dataviz/ →

Flagship Course

Development Economics

The economic theory behind development programmes — what evidence says about poverty, growth, agriculture, credit, and human capital. Provides the sector knowledge that makes MEL findings interpretable.

impactmojo.in/courses/devecon/ →

ImpactMojo 101 Series: This deck is part of the ImpactMojo 101 Series — standalone 100-slide primers covering development concepts at no cost. Other decks in the series cover Climate Essentials, Development Economics, and more. All free, forever, at impactmojo.in

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Gender & MEL Deep Dive

Gender-Responsive MEL: A Practical Toolkit

Gender Analysis Tools for MEL

Tool	What It Measures	When to Use
WEAI (Women's Empowerment in Agriculture Index)	5 domains: decisions, resources, income, leadership, time	Agricultural programmes; baseline and endline
GiHA (Gender in Humanitarian Action)	Gender-specific needs and capacities in emergency contexts	Humanitarian response; rapid assessment
Oxfam Gender at Work framework	Formal/informal rules and individual/social change across two axes	Women's rights and empowerment programmes
Social Norms Assessment (SoNA)	Prevailing social norms affecting women's behaviour; peer expectations	Behaviour change programmes; SBCC evaluation
EMERGE standards	Evidence and Methods for Gender Responsive Evaluation — comprehensive framework	Any evaluation with significant gender dimensions

Time use studies in gender MEL: One of the most neglected but revealing tools for gender analysis. Time use surveys measure how men and women spend their time across productive, reproductive, and care activities. In India, women perform on average 5–6 hours of unpaid care work daily (PLFS data) — invisible in most MEL systems. Programmes that improve women's incomes without reducing their unpaid care burden may not improve wellbeing and may increase time poverty.

Separate data collection for women: In mixed-gender contexts in many parts of rural India, women will not give candid responses in the presence of male family members or field staff. Women enumerators, separate interview spaces, and women-only FGDs consistently produce higher quality data on sensitive topics — income control, mobility, domestic violence experience, contraceptive use — than mixed-setting data collection.

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Evaluation Ethics

Evaluation Ethics: Standards and Responsibilities

Evaluation involves power — the evaluator has power over how findings are framed, which voices are amplified, and what gets reported. Ethical evaluation requires deliberate attention to these power dynamics — not just compliance with research ethics protocols.

Core Ethical Principles in Evaluation

Respect for persons: Treat participants as ends, not means. Informed consent, confidentiality, right to withdraw.
Beneficence: Maximise benefits, minimise harms — including the harm of conducting a poor-quality evaluation that wastes community time and produces useless findings
Justice: Fair selection of participants; representation of marginalised voices; equitable distribution of evaluation burdens and benefits
Independence: Evaluation findings must not be shaped by the interests of the commissioning organisation
Transparency: Methods, data, and funding sources disclosed; limitations acknowledged

The commissioner's influence problem: Evaluations are commissioned and paid for by the organisations whose programmes they assess. This creates systematic pressure — explicit or implicit — to produce positive findings. The most common manifestations: evaluators selected based on likelihood of producing positive findings; draft reports heavily edited before publication; negative findings downplayed or removed. These practices corrupt the evaluation function and should be named and resisted by evaluators.

Institutional Review Boards (IRBs) in India: For research-grade evaluations involving human subjects, IRB (or IEC — Institutional Ethics Committee) approval is required. The Indian Council of Medical Research (ICMR) sets guidelines for biomedical and health research ethics. For social science research, standards are less formalised but ICSSR guidelines apply. Larger international NGOs and academic partners will have their own IRB processes.

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Evidence-Informed Practice

Evidence-Informed Practice: Using Research in Programme Design

Beyond MEL for your own programmes, evidence-informed practice means drawing on the existing sector knowledge base when designing interventions — not starting from scratch. The development research literature is vast; navigating it is a skill that separates evidence-informed programmes from those designed on assumption and institutional habit.

Evidence Portals for Development

3ie Evidence Portal (3ieimpact.org): Systematic reviews and impact evaluations across development sectors. Best for finding rigorous evidence summaries on what works.
Campbell Collaboration (campbellcollaboration.org): Systematic reviews in social welfare, education, crime and justice, international development.
J-PAL Policy Insights (povertyactionlab.org): Evidence summaries from RCTs; policy recommendations; India-specific evidence.
ImpactMojo DevDiscourses: 500+ open-access development papers curated for South Asian practitioners. impactmojo.in

Using evidence practically: Before designing a new programme, ask: (1) What does the evidence say about effective approaches to this problem in similar contexts? (2) What have comparable programmes found about the mechanisms that do and don't work? (3) What are the documented implementation challenges? (4) What outcome levels are realistic given the evidence base? Starting with evidence changes what you design, how you measure it, and what you expect.

The limits of evidence transfer: Evidence from one context does not automatically transfer to another. RCT findings from Kenya, Bangladesh, or Peru give useful signals — but context determines whether mechanisms fire. The question is not "does evidence say this works?" but "does evidence say this works in contexts similar to mine, and what are the key contextual factors?" This is where evidence synthesis and realist evaluation thinking are most valuable.

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Capacity Building

Building MEL Capacity: Individual, Team, and Organisation

MEL capacity building is not just training staff in data collection tools. It requires developing competencies at individual, team, and organisational levels — and changing the systems and culture that determine whether those competencies are used.

Level	What to Build	How
Individual	Technical skills: survey design, data analysis, report writing, evaluation methods	Courses (ImpactMojo MEL); mentoring; peer learning; practice with feedback
Team	Shared understanding of MEL purpose; collective data interpretation; learning culture	Team-based learning events; joint data reviews; after-action reviews; cross-team learning
Organisation	MEL systems; data governance; learning processes; leadership that values evidence	MEL system audit; MEL plan development; leadership engagement; funder conversations about learning

The training-only trap: Sending a MEL officer to a training course does not build organisational MEL capacity. Individual training without supportive systems, leadership buy-in, and time for application rarely changes practice. Sustainable MEL capacity building requires: (1) leadership commitment to using evidence; (2) dedicated time for MEL activities in staff workplans; (3) systems that make using evidence the path of least resistance; (4) funding for MEL that is not treated as overhead.

MEL budgeting: Industry norm for MEL as percentage of programme budget: 5% for small, well-established programmes; 7–10% for medium complex programmes; 12–15% for programmes with high learning ambitions or complex evaluation designs. When MEL is underfunded, it produces incomplete data, overworked staff, and compliance-only outputs. Make the budget case using the cost of poor decisions from missing evidence — not just the cost of the MEL activities themselves.

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Sample Evaluation ToR

Sample Evaluation ToR Structure: Women's Livelihoods Programme Endline

Sections 1–4

1. Background and Programme Description: Organisation; programme area (3 districts, MP); target population (women aged 18–50 from ST/SC households); programme budget (₹4.2Cr, 4 years); theory of change summary; what has been implemented
2. Purpose and Audience: Endline evaluation; primary audiences: organisation's Board and FCDO programme officer; decision to be made: whether to seek Phase 2 funding and at what scale; secondary: state government partnership discussions
3. Evaluation Questions (maximum 5): (1) Did women achieve sustained enterprise operation and income at programme close? (2) What factors distinguish high-performing from low-performing participants? (3) To what extent did the programme contribute to observed income changes, compared to other factors? (4) Were benefits sustained beyond programme support? (5) What implementation factors enabled or constrained outcomes?
4. Evaluation Design: Mixed methods; quantitative survey with comparison group (Phase 2 districts where programme not yet implemented); qualitative component with FGDs and case studies; contribution analysis framework

Sections 5–8

5. Data Requirements: Existing data available: baseline survey (n=420); quarterly monitoring reports; field records; MFI disbursement data. Primary data to collect: endline survey (n=400 treatment, n=200 comparison); 8 FGDs (4 treatment, 4 comparison); 12 KIIs; 6 case studies
6. Ethics: IEC approval from [organisation's IEC]; informed consent protocol for all primary data; women enumerators for all female respondents; data storage protocol; anonymisation before analysis
7. Deliverables and Timeline: Inception report (Week 3); preliminary findings (Week 10); draft evaluation report (Week 14); stakeholder validation workshop (Week 16); final report (Week 18); executive summary in Hindi (Week 19)
8. Budget and Logistics: Evaluation budget ₹18L; includes field team costs, travel, data processing; organisation provides transport support; evaluator responsible for own accommodation

ImpactMojoMEL Basics 101www.impactmojo.in

Supplementary · Final Reflection

A Final Word on MEL and the Sector

The development sector spends enormous resources on MEL — surveys, evaluations, reporting systems, MEL officers — and produces relatively little institutional learning. The gap between MEL investment and MEL value is not primarily a technical problem. It is a cultural and structural one.

The question is not whether we can measure what matters, but whether we are willing to be honest about what the measures show — especially when they show that we are wrong.

ImpactMojo — MEL Basics 101

What Makes the Difference

Organisations where leadership uses evidence to make decisions — not to justify decisions already made
Funders who reward honest reporting on what didn't work — not just what did
MEL practitioners who push back on overclaimed attribution and vague indicators — even when it's professionally uncomfortable
Communities who are given the data, the feedback channels, and the power to assess programmes themselves
Practitioners who see MEL as service to the people their programmes claim to serve — not as compliance with the people who fund them

You already have the tools. The concepts in this deck — Theory of Change, results chains, indicator design, evaluation design, contribution analysis, learning loops — are not advanced technical skills. They are frameworks for clear thinking about what you are trying to do, whether it is happening, and what you are learning. That discipline, applied honestly, is what good MEL is.

ImpactMojoMEL Basics 101www.impactmojo.in

ImpactMojo 101 Series

100 Slides.
Free Forever.

MEL Basics 101 is part of ImpactMojo's free open-access education platform for South Asian development practitioners. PhD-level rigor, zero cost.

MEL Flagship Course Data Visualisation Development Economics

www.impactmojo.in · CC BY-NC-SA 4.0 · Share freely, attribute always