Why Self-Reported Project Data Fails Funders — and What Independent Verification Looks Like in West Africa
Over the last decade, the tools for collecting field data have become remarkably good — and in many cases free. Open-source survey platforms, mobile form builders, geotagged photo workflows: an implementing organisation today can stand up a data-collection system that would have cost a fortune in 2015.
And yet funders keep getting surprised. Milestones reported complete turn out not to be. Sites reported active turn out to be quiet. The dashboard looked healthy right up until it didn't.
The tools are not the problem. The problem is a structural one that no collection tool can fix: when the party doing the work is also the party reporting on it, the data is self-reported — no matter how sophisticated the pipeline that carries it.
Better plumbing, same water
A modern data-collection stack does genuinely useful things. It standardises questions, timestamps submissions, attaches GPS coordinates and photos, and lands everything in a tidy database. Well-known platforms like KoboToolbox do this well, at little or no cost, and are rightly the default across the humanitarian and development sector.
But look at what the stack actually guarantees. It guarantees that someone submitted something, in a structured format, at a recorded time. It does not guarantee that the someone was where they claimed, that the photo shows what it purports to show, or that the status selected reflects the site rather than the pressure the reporter is under. Every guarantee in the pipeline is about format. None is about truth.
There is also a quieter cost that "free" hides. A self-service stack needs devices procured, forms designed, enumerators trained, and a staff member to maintain dashboards and chase submissions. Organisations routinely spend more in staff time running a free tool than a managed service would cost — and at the end of it, the data is still self-reported.
The three questions every funder should ask of a data point
Who captured this — and did they have a stake in the answer?
What evidence accompanies it — and was the evidence captured at the moment of reporting, or attached later?
Could anyone edit this after the fact?
Most reporting systems cannot answer these questions at all, because the information was never recorded. The report says "site active, 60% complete." It does not say — anywhere — whether that is an implementer's claim, a photographed observation, or an independently verified fact. Three radically different levels of reliability, rendered identically.
What honest data looks like: graded evidence
The fix is not more collection. It is making the pedigree of every data point visible and permanent.
In FieldIntel — the platform built by our firm, InfraPro (SL) Limited, in Freetown — every field data point carries an evidence grade fixed at the moment of capture:
Verified — recorded on site by an accredited InfraPro officer during a physical visit.
Photo-backed — reported from the field with photographic evidence captured at submission.
Claim-only — a claim, honestly labelled as a claim.
The grade is applied automatically from the facts of capture — never assigned by judgement, never editable afterwards, by anyone. A funder reading a report sees not just what was reported but how much weight it can bear. Gaps and unverified claims stay visible instead of dissolving into an averaged dashboard.
And when it matters, independence
Graded evidence tells you which data points need checking. Independent verification is the check. An accredited officer with no stake in the implementation visits the site and records what is physically there — and that finding enters the record at the highest grade.
Historically, this kind of independence came only in the form of large third-party monitoring contracts, priced for the largest programmes. Commissioned per visit, it becomes something any funder can do routinely: verify a milestone before a disbursement, spot-check a portfolio quarterly, put eyes on one site that feels wrong.
When self-reporting is enough — and when it isn't
Self-reported data, honestly labelled, is fine for low-stakes operational awareness. It stops being fine when money, compliance or reputation depend on it: disbursement milestones, donor reporting, completion certification, anything a board must stand behind. The rule of thumb is simple: the moment a data point triggers a decision, someone independent should be able to stand behind it — or it should visibly carry the label "unverified claim."
Funders do not need perfect data. They need to know which parts of their data are load-bearing.
See what graded, verifiable reporting looks like
We will walk you through a live evidence record — self-reported, photo-backed and independently verified data side by side, and the reports each produces. Thirty minutes, no obligation.