Several structures, one reality
An organisation with three grants does not have three sets of costs. It has one set of costs and three descriptions of them, each written by a different funder, each with its own headings, dates and conditions.
That is a manageable arrangement and it becomes unmanageable in one specific way: when each description is maintained separately. Three spreadsheets, three people, three quarter-ends — and no single place where you can see that a cost has been claimed once and only once.
Everything on this page follows from one principle. Keep a single record of what was spent, and produce each funder’s view from it. Never the other way round.
What that buys is not tidiness. It is that adding a fourth grant costs a few hours rather than a third of someone’s job.
The shape of the problem
Worth laying out, because the difficulty is not any single grant — it is that the differences between them are all in different dimensions.
| What differs per grant | Consequence |
|---|---|
| Budget headings | The same cost is named differently in each claim |
| Reporting periods | Quarter ends land in three different weeks |
| Date convention | One claims on invoice date, another on payment |
| Eligible cost lists | A cost fine for one is excluded by another |
| Evidence expectations | Different levels of documentation |
| Overhead treatment | A percentage here, an agreed list there |
Six differences and none of them can be resolved by picking one and applying it everywhere. Each grant genuinely has its own rules, and your job is not to reconcile them but to keep them from contaminating each other.
Which is exactly what a single underlying record does. The costs are recorded once, and each funder’s rules are applied as a view over them rather than baked into the data.
The thing that must never happen
Claiming the same cost against two funders. It is the most serious error in this area and the one least likely to be caught by any ordinary control.
It almost never involves anyone doing anything wrong. Two programmes, two people, two claims prepared from the same folder of invoices, no shared record of what went where. Both claims are individually correct and defensible. Together they claim the same £4,000 twice.
Nothing in either claim reveals it. The invoice is real, the amount is right, the coding is plausible, and neither funder sees the other’s schedule. It surfaces at a verification visit, or when the two funders happen to be the same body under different programmes — and by then it looks considerably worse than the administrative accident it was.
Two checks prevent it entirely, and both need every allocation in one place. Does this document appear against more than one programme? — which is often perfectly legitimate. Do the allocated shares sum to more than the document? — which never is.
The second check is the one that matters and it is arithmetic rather than judgement. The mechanics are described on eligible cost screening.
One table, several views
Concretely: every cost is a row, and the row carries what it needs for any funder to be served from it.
| Column | Why it is there |
|---|---|
| Supplier and document number | Identity — the basis of every duplicate check |
| All three dates | Each funder names a different one |
| Net, VAT, total | Some claim gross, some net |
| Programme | Which grant, or none |
| Share and basis | For anything split, in words |
| Funder heading | Per programme, from the mapping |
| Source file and page | So any figure leads back to its evidence |
The second row is the one people economise on and regret. Keeping all three dates rather than the one you happen to need costs nothing at capture and makes each funder’s period rule a filter rather than a rework.
From that table, a claim is a filter and a grouping. So is a board report, so is an annual return, so is the answer to a query six months later. The work of producing any of them collapses to minutes because none of it involves going back to documents.
Attaching the programme and heading to each row uses the same mechanism as any business dimension, covered on dimension tagging.
The unfunded core, and being honest about it
Most organisations running funded programmes also do things nobody funds: the finance function, the premises, the person who answers the phone, the work between grants.
Those costs are real and they have to sit somewhere. Where they sit is a decision with consequences, and it is worth making deliberately rather than by default.
If your agreements allow overhead recovery — a percentage, or a defined list — apply it as the agreement describes and keep the calculation visible. This is the cleanest case and the least common.
If they do not, those costs stay unallocated, and the total sitting outside all programmes is a number worth watching. It is the true cost of running the organisation, and in many funded bodies nobody has ever seen it as a single figure.
What to avoid is the middle path: spreading unfunded costs across programmes by a basis invented to make the books look complete. It is not claimable, it distorts every programme it touches, and it cannot be explained to anyone who asks.
Seeing that unallocated total plainly is often the most useful by-product of this whole exercise. It is the number that informs what your next funding application needs to cover.
The monthly routine
1 · Capture the documents
Everything that arrived, from everyone. Photographs of receipts included — those are the ones that vanish.
2 · Extract in one batch
Up to 100 files at once, with the sum of lines checked against each document total.
3 · Assign the programme
Rules for the suppliers that always belong to one; a decision for the rest.
4 · Check the shares
Anything split must sum to the document. This is the check that prevents double-claiming.
5 · Ask while it is fresh
The handful of costs nobody can place. Asking now takes a minute; asking in April is archaeology.
6 · Look at the unallocated total
Rising month on month means the rules are not keeping up.
An hour, once the rules exist. Note that no step in it is claim preparation — claims are produced from this table when each funder’s deadline arrives, and by then there is nothing left to assemble.
That separation is the whole point. Monthly work that is not tied to any funder’s calendar is work that happens; work scheduled for the week before three different deadlines is work that gets compressed.
Three sizes of organisation
One person, one or two grants
The same person delivers, spends and reports. The advantage is that the documents and the knowledge are in the same head; the risk is that delivery always feels more urgent than administration, and it usually is.
What works here is fifteen minutes a fortnight rather than a day per quarter. Small enough to survive a busy month, which is the only property that matters at this scale.
A small team, two to four grants
The size where double-claiming becomes a genuine risk, because more than one person now touches the costs and nobody holds the whole picture.
One person owns the table — not necessarily doing all the coding, but responsible for it being one table. An hour a month, and the same person each time so the rules and the bases stay consistent.
A finance function, five or more
Usually there is a system with dimensions already, and the gap is not the reporting but the input: the documents arriving as PDFs and photographs that someone keys in.
That is the part that automates, and the output feeds the system you already have. At this scale the duplicate check should be continuous rather than monthly, because the volume makes a quarterly review impractical.
Across all three, the constraint is identical: the routine has to be small enough to survive a difficult month. A process that requires a good month is a process that runs about half the time.
The part this does not cover
Staff costs are usually the largest heading in a funded budget, and nothing here touches them. They come from payroll and timesheets, and the apportionment between programmes follows your funder’s rules rather than any general method.
What this covers is everything the organisation buys — which is where the documents are, where the evidence requests land, and where the queries come from. It is the minority of the money and the majority of the administration.
Say which one a report contains. “Non-staff costs by programme” is honest and useful; presenting the same figures as full programme cost is neither, and it is how these exercises lose the confidence of the people who have to act on them.
Where you do have timesheet data, it joins the same table as another category of row. The structure does not change; only the source does.
Adding a new grant
If the underlying record is right, a new grant costs a few hours once rather than a permanent increase in workload. Four things to set up, in the first fortnight.
The heading mapping — their budget lines against your categories, with any disputed cases decided in writing with the funder.
The period rule — which date counts, what the reporting windows are, and how late costs are handled.
The rule set — the date window and the exclusions, starting small and growing as questions get answered.
The programme code — one value, never reused, added to the table so allocation can start on day one.
Do all four before the first cost is incurred if you possibly can. A programme that runs for six weeks before anyone sets up its coding produces six weeks of costs that have to be reconstructed, and that reconstruction is the most expensive kind of work in this whole area.
The full setup sequence is in the grant reporting guide.
When a grant ends
A programme stops delivering on one date and stops accepting costs on a later one, and treating those as the same day is how final claims end up incomplete.
Suppliers invoice on their own schedule, so costs belonging to the last weeks arrive over the following month or two. Hold the programme open in your own records past the final claim and look at what lands.
Two other things to get right at the end. Credit notes arriving after a final claim reduce what you were entitled to, and finding one yourself is a very different conversation from a funder finding it. And retention obligations usually outlive the programme by years — check what your agreement requires before any clear-out or system migration, because migrations routinely carry transactions and leave documents behind.
Finally, keep the working table after the grant closes. Verification requests can arrive long after a programme has finished, and the table is what makes answering them a filter rather than an excavation — the standard is on the evidence pack.
The board asks a different question
Funders ask about their own grant. A board asks about the organisation, and the same table answers both — but only if it was built with the second question in mind rather than assembled per funder.
What a board usually wants is three things that no individual claim contains. What proportion of our costs are covered by funding, which is the unallocated total seen from the other direction. Which programmes are on profile, which needs spend against budget across all of them at once. And what happens when the largest grant ends, which needs the first two together.
None of those can be produced from three separate claim spreadsheets without a reconciliation exercise, and that exercise is why board reporting in funded organisations so often lags a quarter behind reality.
From one table it is a grouping. Same rows, grouped by programme rather than by funder heading, with the unallocated costs shown as their own line rather than hidden.
That last detail matters more than it sounds. An organisation that never sees its unfunded costs as a single figure tends to under-ask in funding applications, because nobody has quantified what running the place actually costs — and the figure is usually larger than anyone in the room expects.
Comparing those figures across quarters is what turns them into something a board can act on rather than acknowledge; the mechanism is on period comparison.
Four objections
“Our grants are too different to share a table.”
The costs are not different — the descriptions of them are. A shared table with a programme column handles arbitrarily different funders; separate tables are what makes them incompatible.
“We have never double-claimed.”
Quite possibly true, and unverifiable without the check. The point of the arithmetic is that it is cheap and it makes the claim provable rather than merely believed.
“We don't have the capacity for a monthly routine.”
The monthly version is less total time than the quarterly one, because it excludes reconstructing what each cost was for. The constraint is attention rather than hours.
“Our funders don't ask for this detail.”
Until one does, usually at a verification visit or a final claim. The detail is cheap to keep continuously and expensive to produce on request.
Four signs the arrangement is not holding
Worth checking against your own situation, because each of these is visible long before it becomes a problem a funder notices.
Nobody can say what the unallocated total is. If the answer takes more than a minute, the costs are not in one place, and the double-claiming check cannot be run either.
Each claim starts by opening a folder of documents. That means the extraction is happening at claim time rather than continuously, and the cost per document is several times what it should be.
Two people give different answers about the same supplier. The allocation rules exist in individual heads rather than in a record, and the claims will diverge in ways nobody detects.
A new grant felt like it added a job. Adding a programme should cost a few hours of setup. If it added ongoing work proportional to the first one, the underlying record is being maintained per funder rather than once.
None of the four requires an audit to detect. Each is a question you can ask in a meeting, and the hesitation before the answer is usually more informative than the answer itself.
Where to start
Not with a system, and not with all your grants at once.
Take last month’s documents — all of them, across every programme — extract them into one table, and add a programme column. Then run the one check that matters: do the allocated shares sum to each document?
If they do, you have confirmed something valuable cheaply and you now have the table. If they do not, you have found the thing this exercise exists to find, in about an hour, and while it is still correctable.
Then do the same next month. The second month is where it becomes a habit, and by the third the claims are a filter over a table rather than a fortnight of work.