What a grant report actually consists of
Two documents that have to describe the same quarter: a financial schedule and a narrative. They are usually written by different people, at different times, from different material — and the most common thing a funder queries is that they disagree.
The financial side is a translation: your costs, re-grouped into the funder’s budget headings, restricted to their period, screened against their conditions, backed by documents.
The narrative side describes what the money did. It is not this guide’s subject, with one exception that gets its own step: it has to be consistent with the numbers, and making that true is a mechanical task rather than a writing one.
What follows is eight steps in the order that makes each subsequent report cheaper. The first three are decisions you make once. The next four are the quarterly work. The eighth is the check nobody schedules.
Before you start: four things out of the agreement
Somebody has to read the agreement properly, once. Twenty minutes, with four things written on one page afterwards.
The period — both the overall grant period and the reporting periods inside it, which are not the same thing and get confused constantly.
The budget table — headings and amounts, exactly as the funder writes them, including any that are zero.
The eligible and ineligible cost lists — and any threshold above which extra evidence or prior approval is required.
What has to accompany a claim — a schedule only, or documents; a certification, and by whom.
Anything ambiguous goes on a list to ask about, and the asking happens now rather than at the first claim. A funder answering a question in month one is routine; the same question in month eleven, with a claim attached, reads as a problem.
Keep the answers in writing next to the agreement. An emailed clarification is worth considerably more at audit than a reasonable interpretation you arrived at alone.
Step 1 · Build the heading mapping
The funder’s headings against your own accounts, as a table, written once. This is the single highest-return twenty minutes in grant reporting and it is the step most often skipped because it feels obvious while you are doing it.
It is not obvious three months later, and it is not obvious at all to whoever does the claim after you. Two people mapping the same costs independently produce two different claims, and neither is wrong — which is exactly why the mapping has to be a record rather than a recollection.
Write three columns: their heading, your accounts or cost categories that feed it, and a note on anything you had to decide. The third column is the valuable one.
Where a mapping is genuinely ambiguous — is a freelance facilitator direct delivery or a staff cost? — decide it with the funder rather than internally. It takes one email and it removes an argument that would otherwise recur every quarter for the life of the grant.
Attaching the heading to each extracted row uses the same mechanism as any other business dimension; that is covered on dimension tagging.
Step 2 · Fix the date convention
Every cost has at least three dates: when the work happened, when it was invoiced, and when it was paid. Your agreement names one of them, or should.
Write down which one, and write down what you do with the awkward cases — a pro forma, a deposit, a subscription spanning the boundary, a credit note against a prior period. Those four cover most of what will come up.
The reason this deserves a step of its own is that drift is invisible. Nobody decides to change convention; someone new prepares a claim, uses the date that seems natural, and a cost is either claimed twice or never. Neither shows up in any total.
One habit makes the whole thing safe: keep a record of what you claimed at each boundary. “Did this one go in last time?” is asked every quarter, and it should be answerable in seconds rather than from memory.
Step 3 · Write the screening rules
Turn the conditions from the agreement into rules that can run over rows: the date window, the excluded categories, the thresholds, the caps.
Start with two — a date window and a short exclusion list — because those catch most of what gets caught. Add the rest as your funder answers questions, so that the rule set grows out of real cases rather than out of an attempt to anticipate everything.
Record each rule with what it checks, where in the agreement it comes from, and when it was decided. The middle field is what lets a successor trust a rule they did not write; the third is what lets you defend a claim prepared before a clarification arrived.
Keep one rule set per grant. A single organisation-wide list is tempting and breaks the first time two agreements disagree, which is most of the time. The mechanics are on eligible cost screening.
Step 4 · Collect documents as you go
This is where the three weeks go, and it is the step with the highest ratio of pain to difficulty.
At claim time the figures are usually available and the documents are not. They are in an inbox, in a shared drive, in the downloads folder of someone who has since left. Every one of them takes between two minutes and an afternoon to find.
The fix is unglamorous: capture the document when the cost arrives. One folder per grant per period, or better, extract the document at that point so the row and the file are already connected.
Doing this weekly costs a few minutes. Doing it once at the end of a quarter costs days, and the difference is not effort — it is that at the end you are also reconstructing which cost belonged where, from documents you are seeing for the first time.
Practically: process a batch each week or fortnight, up to a hundred files at a time, and let every row carry the file and page it came from. What you end up with at quarter end is a table rather than a task. The batch mechanics are on merging documents into one export.
Step 5 · Allocate the shared costs
Costs that serve more than one programme, or a programme and your unfunded work. Three rules keep this defensible.
Split at line level where the lines allow it. A line-level split is a fact about the document. An invoice-level split is a judgement, and judgements need explaining.
Record the basis in words.“Split by staff hours on each programme, per the timesheet summary” can be discussed. A bare percentage cannot be defended and should not be attempted.
Make the shares sum to the document. Check it. The catastrophic failure in this area is claiming 60% to one funder and 60% to another, and it happens when two people work from the same folder without a shared record.
There is a fourth option that is legitimate and underused: leaving a cost unallocated on purpose. A cost that genuinely serves the whole organisation should say so rather than being spread by a basis invented to avoid an awkward blank.
Step 6 · Assemble the schedule
With the first five steps done this is mechanical, which is the point of doing them.
Filter the period on your date convention. Run the screening rules and look at what they flag. Group by heading. Total each heading and compare it to the budget. Export with the source document on every row.
Two things to check before it goes anywhere. First, that every heading with spend is one the funder recognises — a heading that appears in your schedule and not in their budget is a mapping error that will come straight back. Second, that no heading has quietly passed its cap; if one has, that is a conversation to have before the claim rather than inside it.
Keep the working table, not just the submitted summary. When a query arrives in six weeks, the summary answers nothing and the table answers everything.
Step 7 · Reconcile to your own ledger
The step that gets skipped when a deadline is close, and the one that catches the errors worth catching.
Take the total you are about to claim and find it in your own books. If the two do not agree, one of three things is true: a cost is in the claim and not in the ledger, a cost is in the ledger and not in the claim, or the same cost is in both under different amounts.
All three are worth knowing about before a funder finds them, and the third is the one that indicates a real problem rather than a timing difference.
Expect some difference and understand why. Timing conventions, accruals, and costs coded to the grant in your books but screened out of the claim all produce legitimate gaps. What you want is not zero difference but a difference you can explain in one sentence.
Write that sentence down and keep it with the claim. It is the first thing you will want six months later, and it takes thirty seconds now.
Step 8 · Make the narrative agree with the numbers
The financial and narrative reports are usually written separately, often by people who do not compare them. Funders do compare them, and it is the most common source of a query.
Three checks, ten minutes. Does every activity described in the narrative carry cost somewhere in the schedule? Does every heading with significant spend get mentioned in the narrative? Do the two describe the same period — including the same treatment of anything that slipped?
The mismatches are rarely dishonest. A narrative written from the plan describes what was supposed to happen; a schedule built from documents describes what did. Where they diverge, the divergence is usually the most interesting thing in the report and worth saying out loud.
A funder reading “the second workshop moved to next quarter, so delivery costs are below profile” needs no follow-up. The same situation with no explanation generates an email, a reply, and two weeks.
A worked example
A small organisation with one grant, a quarter of costs, and the method above already in place.
| Step | What happens | Time |
|---|---|---|
| Documents | Already captured weekly — nothing to find | 0 |
| Extract the last batch | Two weeks of invoices, one pass | 10 min |
| Apply the date rule | Four rows near the boundary, reviewed | 10 min |
| Screening | Six flags: five included, one moved | 20 min |
| Shared costs | Two suppliers split on a recorded basis | 15 min |
| Assemble and check headings | One heading at 91% of cap — noted | 20 min |
| Reconcile to the ledger | £340 difference, explained by an accrual | 20 min |
| Narrative check | Workshop slipped — added a sentence | 10 min |
Under two hours, and the largest single item is a reconciliation that exists to catch problems rather than to produce anything.
Compare that with the same quarter done without the method: the documents step alone is two days, because it includes working out what each one was for.
The quarterly rhythm
What makes the difference is not doing more work but doing it at different times.
| When | What you do | Why then |
|---|---|---|
| Weekly | Capture and extract documents | They are findable now and not later |
| Monthly | Check headings against budget | A cap breach is a decision, not a surprise |
| At quarter end | Assemble, screen, reconcile | Everything is already in place |
| Two weeks after | Look at what arrived late | Suppliers bill on their schedule |
The second row is the one that changes outcomes rather than effort. A heading tracked monthly gives you a quarter’s notice that it is running hot, and a quarter is enough time to move something, request a virement or slow the spending. The same information at claim time is just bad news.
When a funder queries something
It will happen, and it is usually routine. How it goes depends entirely on whether you kept the working table.
“What is in this heading?” — a filter and a reply. Minutes, if the rows exist.
“Can we see the invoice for this?” — following a reference, if each row carries its document. A search through folders otherwise.
“Why is this split 60/40?” — reading back the recorded basis. Impossible to answer convincingly if the basis was a judgement nobody wrote down.
“Why does this differ from your accounts?” — the sentence you wrote at step 7, retrieved rather than reconstructed.
Answer quickly and completely even when the query is trivial. A funder who gets a clear answer in a day asks fewer questions next time; one who waits two weeks for a partial answer starts looking at everything else.
The final claim, which is not like the others
Every claim before the last one has a next one to correct it in. The final claim does not, and that changes how it should be prepared.
Hold it longer than feels comfortable. Costs belonging to the last weeks of a programme arrive over the following month or two. Submitting the day the programme ends means submitting before you know what it cost. Where a funder allows a window for the final claim, use all of it.
Check every heading against its budget. Underspend on one heading and overspend on another can often be reconciled by agreement at this point, and almost never after submission. This is the last moment that conversation is available.
Look for credit notes deliberately. A credit against a cost claimed in an earlier period reduces what you were entitled to. Finding it yourself and adjusting is routine; a funder finding it in a final verification is not.
Expect a fuller evidence request. Final claims attract more scrutiny than interim ones, and often a full pack rather than a sample. An organisation that collected documents throughout has nothing to do; one that did not has a project.
Write the closing note. One page: what was claimed in total, how it compares to the budget, what moved between headings and why, and anything you know is unusual. It costs half an hour and it answers most of what would otherwise arrive as questions.
If you are audited or verified
A verification visit or an audit of grant expenditure is not an accusation. It is routine assurance, and most of them are uneventful for organisations that kept their material in order.
What is examined is narrower than people fear. Typically: a sample of claimed costs traced to documents, the allocation basis for anything shared, whether the period treatment was consistent, and whether anything claimed appears to be excluded by the agreement.
Notice that all four are answerable from the working table if it exists, and from nowhere at all if it does not. That is the whole reason step 6 said to keep the rows rather than only the submitted summary.
Three habits make a visit go well. Answer what was asked rather than volunteering a quarter when three lines were queried. Disclose gaps yourself — a missing receipt you name is handled very differently from one that is found. Have one person answering, so that two people do not give two accounts of the same allocation.
And treat the findings as rules. Whatever an auditor raises is something the next claim should screen for automatically, which turns a one-off inconvenience into a permanent improvement. What evidence is expected for each cost type is set out on the evidence pack.
Common mistakes
Leaving document collection to claim time
The largest cost in grant reporting, and the easiest to remove. Capture as costs arrive.
Redoing the heading mapping each quarter
Two people reason differently from the same costs and produce two different claims.
Letting the date convention drift
Nobody decides to change it; someone new just uses the date that looks natural.
Interpreting an ambiguous condition alone
One email to the funder is worth more at audit than the most reasonable interpretation.
Skipping the reconciliation when time is short
It exists precisely for the quarters where time is short.
Writing the narrative from the plan
The plan describes what was meant to happen; the schedule describes what did. Funders compare them.
Submitting a total instead of keeping the rows
The summary answers nothing when a query arrives six weeks later.
Handing it over
Grant reporting outlives the people who set it up. The most common way a well-run process collapses is not a decision to stop but a change of staff, with the method leaving in someone’s head.
Four things have to exist in writing for a successor to continue rather than reinvent: the heading mapping including the cases you had to decide, the date convention with its awkward cases, the screening rules with their source and date, and the record of what has been allocated where.
The fourth is the one people underestimate. Without it a successor cannot tell whether a cost has already been claimed, and the safest-looking option — claiming it again — is the one error that matters most.
Add a one-page note on the relationship itself: who the contact is, what has been asked and answered, and anything the funder has clarified that is not in the agreement. That page is worth more than any of the tables, because it is the only part that exists nowhere else.
Checklist
Period, budget table, cost lists and claim requirements written on one page.
Ambiguities asked about in writing, answers kept with the agreement.
Heading mapping recorded, including the cases you had to decide.
Date convention chosen, with the awkward cases written down.
Screening rules recorded — what, from where, decided when.
Documents captured weekly rather than hunted at quarter end.
Shared costs split at line level, basis recorded in words.
Shares checked to sum to the document, across all programmes.
Headings checked against budget monthly, not at claim time.
Claim reconciled to the ledger, with the difference explained in a sentence.
Narrative and schedule compared before submission.
Working table kept, not just the submitted summary.