The AJ Global Blog —
Evidence, Governance & Field Perspective

For most of the last two decades, monitoring and evaluation in donor-funded programmes has revolved around a familiar artefact: the log-frame. Indicators are set at design stage, data is collected on a quarterly cycle, and results are reported upward at the end of the reporting period — often long after the moment to act on them has passed. In fast-moving, fragile, and resource-constrained environments across Nigeria and the wider Sahel, that lag is no longer a minor inefficiency. It is a structural risk to programme quality and donor confidence.

An adaptive learning system flips this sequence. Rather than treating monitoring as a compliance exercise that produces a report, it treats monitoring data as a live input into programme management — reviewed on a cadence short enough to inform decisions while they still matter. In our engagements supporting donor-aligned MEL frameworks across Northeast Nigeria, we have seen the same pattern repeatedly: programmes that build structured feedback loops into their design adapt faster, spend more efficiently, and produce stronger evidence of impact at endline.

What an Adaptive Learning System Actually Requires

Building adaptive capability is not simply a matter of collecting data more frequently. It requires four things working together:

• A theory of change that names its assumptions. Adaptive management only works if you know which assumptions, if proven wrong, should trigger a course correction — and which are peripheral.
• Indicators that are sensitive enough to detect change early. Many programme indicators are designed to demonstrate impact at endline, not to detect emerging problems mid-cycle. Adaptive systems need a parallel set of "early-warning" process indicators.
• A review rhythm that matches the operating environment. In conflict-affected or rapidly evolving contexts, a quarterly review cycle may already be too slow. Many of our clients have moved to monthly structured learning reviews for high-risk activities.
• Decision authority at the point of learning. Data that surfaces insight but cannot translate into a decision — because authority sits three approval layers away — produces frustration, not adaptation. Effective systems push proportionate decision authority closer to the field.

Embedding Learning Without Compromising Donor Compliance

A common concern we hear from programme teams is that "going adaptive" risks creating tension with donor reporting requirements that expect consistency against pre-agreed indicators. In practice, the two are complementary rather than contradictory. Donor-aligned MEL frameworks — the kind major bilateral and multilateral donors expect — already build in space for documented, justified adaptation: variance analysis, revised work plans, and learning briefs that explain why and how a programme adjusted its approach.

The institutions that manage this well treat their official indicators as the "spine" of accountability, while running a complementary internal learning system that surfaces issues earlier, tests course corrections at small scale, and feeds validated changes back into the formal reporting structure — with a clear audit trail showing the evidence behind each adjustment. This is precisely the kind of system design our MEL & Programme Support Unit builds with client teams: one that satisfies donor assurance requirements while giving programme leadership the real-time intelligence to manage delivery actively rather than retrospectively.

"The programmes that adapt fastest are not the ones with the most data — they are the ones whose data reaches a decision-maker while there is still time to decide."

A Practical Starting Point

Organisations do not need to redesign their entire MEL architecture to begin building adaptive capability. We typically recommend starting with a single high-risk activity stream: define two or three early-warning indicators, set a short review cadence (monthly or even bi-weekly), assign clear decision authority for that stream, and document each adaptation with a brief rationale. Once that cycle proves itself — usually within one or two reporting quarters — the model can be extended across the wider programme architecture with donor visibility built in from the outset.

Done well, adaptive learning systems do more than improve programme performance. They build the kind of institutional memory and documented decision discipline that donors increasingly look for when assessing an organisation's readiness for larger, more complex mandates.

In Northeast Nigeria, climate change does not arrive as a stand-alone shock. It arrives layered onto displacement, conflict, weakened public services, and already-stretched livelihoods — compounding vulnerabilities that were severe before the rains became less predictable and the dry seasons longer. Designing climate adaptation programming for this context requires more than importing global resilience frameworks. It requires evidence generated inside the context itself.

Our field research across Borno, Adamawa, and Yobe States consistently surfaces the same insight: communities are not passive recipients of climate risk. They are already adapting — shifting planting calendars, diversifying income sources, renegotiating access to grazing and farming land — often without external support, and sometimes in ways that create new tensions. Programming that ignores these existing adaptive strategies risks either duplicating what communities already do or, worse, undermining fragile coping mechanisms that took years to build.

Three Evidence Gaps That Undermine Climate Programming

• Hyper-local climate data. National and regional climate models rarely capture the variation that matters at LGA or ward level — where rainfall failure in one community can coincide with flooding twenty kilometres away. Programmes that rely solely on macro-level data risk targeting the wrong interventions to the wrong places.
• Conflict-climate interaction effects. Climate stress can intensify resource-based conflict — over land, water points, and grazing routes — in ways that purely environmental assessments do not capture. Effective adaptation programming requires joint climate-conflict analysis, not two parallel assessments run in isolation.
• Gendered and displacement-specific vulnerability data. Displaced households, female-headed households, and host communities each experience climate stress differently — and respond to interventions differently. Aggregate vulnerability indices often mask these differences, leading to designs that work for the "average" household but fail the households most at risk.

What Evidence-Based Adaptation Programming Looks Like in Practice

Building climate adaptation programmes on credible local evidence means investing — early, and deliberately — in baseline research that combines household-level vulnerability assessment, community consultation on existing coping strategies, and joint analysis of climate and conflict dynamics. From there, programme design can move from generic resilience activities ("provide drought-resistant seed varieties") toward context-specific interventions co-designed with the communities who will use them, and tested at small scale before wider rollout.

"The most resilient programmes are not the ones that introduce the most new technology — they are the ones that strengthen what communities are already doing right, and remove the barriers that stop those strategies from scaling."

Equally important is building monitoring systems that can detect whether an adaptation intervention is having unintended effects — for example, whether a new water access point is shifting grazing patterns in ways that create friction with neighbouring communities. This is where strong MEL design and conflict-sensitive research design intersect directly with climate programming: an intervention that looks successful against its primary indicator can simultaneously be creating new risks that only a well-designed monitoring system will catch in time.

Why This Matters for Donors and Implementing Partners

Donors increasingly expect climate adaptation investments to demonstrate not just activity completion, but durable change in community resilience — change that can withstand the next shock. That standard of evidence cannot be retrofitted at endline; it has to be designed into the programme from the start, with baselines that are rigorous enough to support a credible before-and-after comparison and monitoring systems sensitive enough to detect both intended and unintended effects along the way.

For implementing partners operating in the Lake Chad Basin and similar contexts, the organisations that invest early in this kind of evidence base are the ones best positioned to demonstrate impact convincingly — and to compete credibly for the next generation of climate financing.

Few moments test an NGO's institutional credibility as directly as a donor compliance audit. The organisations that pass through these reviews smoothly are rarely the ones that scrambled to prepare documentation in the weeks before the audit team arrived. They are the ones that built internal control systems strong enough that an audit simply confirms what good financial governance already produces day to day.

Across our work supporting NGOs, government partners, and programme implementers with audit readiness and grant compliance, we consistently see the same root cause behind compliance failures: not fraud, and rarely incompetence, but the absence of documented systems that translate good intentions into consistent practice. Internal controls exist precisely to close that gap — turning "we are careful with donor funds" into a demonstrable, auditable system.

The Core Components of a Donor-Ready Control System

• Segregation of duties. No single individual should be able to initiate, approve, and record a financial transaction. Even small organisations can build proportionate segregation by separating these functions across two or three people and documenting the approval chain.
• Documented procurement procedures. Donor audits scrutinise procurement closely because it is where the largest sums move and where conflicts of interest most often surface. A clear, written procurement policy — with thresholds for competitive bidding, vendor vetting, and conflict-of-interest declarations — is one of the highest-value documents an organisation can produce.
• Real-time financial reconciliation. Waiting until year-end to reconcile bank statements, cash advances, and ledger entries all but guarantees that errors compound undetected. Monthly reconciliation, reviewed by someone outside the transaction chain, is a foundational control.
• Asset registers and field-office cash controls. For organisations operating in field locations — particularly in Northeast Nigeria and similar contexts — physical asset tracking and cash-handling procedures at the field level are frequently the weakest link, and frequently the first thing auditors test.
• An internal audit or review function. Even a lightweight, periodic internal review — conducted by someone independent of day-to-day operations — catches the majority of issues long before an external auditor does, and demonstrates a culture of accountability that donors notice.

"Donors are not looking for organisations that have never made an error. They are looking for organisations that would catch the error themselves — and could show exactly how."

Building the System Without Overburdening a Lean Team

Many NGOs — particularly smaller and growing organisations — worry that robust internal controls require a finance department they cannot yet afford. In our experience, the opposite is usually true: well-designed controls reduce the burden on lean teams by making each step predictable, documented, and easy to delegate. The investment is front-loaded — in policy design, templates, and a short period of guided implementation — after which the system runs with modest ongoing effort.

We typically recommend organisations start by mapping their current financial process end to end, identifying the two or three points of greatest exposure (often procurement and field cash management), and building documented controls there first. From that foundation, the wider control framework — covering budgeting, reporting, asset management, and internal review — can be built out in a structured sequence aligned to the organisation's grant cycle and audit calendar.

The Pay-Off: Donor Confidence and Funding Continuity

Strong internal controls do more than help an organisation pass its next audit. They are now a key factor in how donors assess organisational risk when deciding whether to expand a partnership, increase grant size, or extend direct funding rather than channelling support through intermediaries. Organisations that can demonstrate a mature, documented control environment are simply easier — and lower-risk — to fund. In a funding landscape where donor due diligence continues to tighten, that demonstrable maturity is becoming a genuine competitive advantage.

Research conducted in fragile and conflict-affected settings carries a responsibility that research in stable environments does not: the process of generating evidence can itself create risk — for respondents, for enumerators, and for the communities being studied. In Northeast Nigeria and comparable contexts across the Sahel and Lake Chad Basin, a research design that would be perfectly sound elsewhere can, without adaptation, expose participants to retaliation, deepen community tensions, or produce data that is simply not credible because people could not safely tell the truth.

Conflict-sensitive research design starts from a different premise than conventional methodology: that "do no harm" is not a separate ethical add-on to be checked at the end, but a design principle that shapes every methodological choice — from sampling strategy to question wording to how findings are eventually shared.

Five Principles That Shape Conflict-Sensitive Design

• Context analysis before instrument design. Understanding local power dynamics, lines of tension, and who is — and is not — safe to be seen speaking to researchers must come before a single survey question is drafted. This analysis directly shapes who can be interviewed, where, by whom, and how.
• Locally-grounded enumerator selection and training. Enumerators who understand local context can read situations that an external team would miss — recognising when a location, a question, or a respondent grouping has become unsafe, and adapting in real time. Their safety briefing must be as rigorous as their data-collection training.
• Sampling strategies that avoid creating visible targets. Highly visible data-collection events — large gatherings, conspicuous vehicles, lists of named participants — can mark individuals or locations for unwanted attention. Conflict-sensitive sampling distributes visibility and protects identifying information at the point of collection, not just in storage.
• Question design that respects what people can safely say. Direct questions about armed actors, security incidents, or political allegiances can put respondents at risk regardless of how the data is later anonymised — because the asking itself is observed. Skilled question design finds ways to generate the needed insight without requiring respondents to make unsafe direct statements.
• Real-time risk monitoring, not just ethical sign-off at design stage. Contexts shift quickly. A location that was safe for data collection in week one may not be in week three. Conflict-sensitive research builds in continuous risk review — with clear authority to pause, adapt, or relocate activity — rather than relying solely on an ethics approval secured before fieldwork began.

"In fragile contexts, the quality of your data and the safety of your respondents are not separate concerns to be balanced against each other — they rise and fall together."

Why This Produces Better Evidence, Not Just Safer Research

It is a common misconception that conflict sensitivity is primarily a constraint on what research can achieve — a set of limitations layered onto an otherwise "ideal" design. In our experience, the opposite is closer to the truth. Research that fails to account for context produces respondents who are guarded, selective, or simply unwilling to participate honestly — generating data that looks complete on paper but is hollow in substance. Research that earns trust by demonstrating genuine understanding of local risk produces richer, more candid, and ultimately more useful evidence.

This is precisely why donor agencies operating in fragile and conflict-affected states increasingly require evidence that proposed research methodologies have been conflict-sensitivity reviewed before approval — not as a procedural formality, but because they have seen, repeatedly, the difference it makes to both participant safety and evidence quality.

Building This Capability Into Your Organisation

Organisations planning research, baseline studies, or evaluations in fragile settings benefit from building conflict-sensitivity review into their research design process as a standing requirement — not an occasional specialist consultation. That means training research leads to conduct basic context and risk analysis themselves, developing standard protocols for enumerator safety and real-time risk escalation, and building review checkpoints into every research timeline, from design through to dissemination of findings.

Done consistently, this approach does more than protect participants — it builds an organisation's reputation as a research partner that communities, donors, and government counterparts alike can trust to operate responsibly in the most demanding environments.