Two halves, and one of them is harder
A recovery report sets what a firm put into client work against what it got back for it. Time on one side, expenses on the other, billed against incurred.
The two halves arrive from completely different places. Time is entered by the person doing the work, into a system built for it, usually the same week. It comes out as structured data because it went in as structured data.
Expenses arrive from third parties, at unpredictable intervals, in whatever format each of them uses — an emailed PDF, a portal download, a photograph of a till receipt, a monthly statement covering forty matters. None of it is data until somebody reads it.
So the report gets built with a precise time half and an approximate expense half, and the approximation is not random: it is biased downwards, because the costs missing from it are exactly the ones that were never captured. A firm reporting 94% expense recovery on a base that is missing a tenth of its costs is reporting a number that means nothing at all.
What we do not do
We do not record time
There is no timer, no timesheet and no matter clock here. Time lives in your practice management system, and this page is about the other column.
We do not raise bills
What was billed comes from whatever raises your bills. We supply what was spent, which is the side that is usually incomplete.
We do not set a target
Any published benchmark for recovery is meaningless without knowing your engagement terms. The useful comparison is your own number over time.
We do not decide what should have been recovered
That is a judgement made against your engagement letter. We make sure the cost is in the report so the judgement is possible.
The question it is supposed to answer
Not “how profitable are we”, which is a different and larger question, but something much narrower: of the money we laid out on clients' behalf, how much came back?
That number is interesting in itself and far more interesting when it is split. A firm at 88% overall could be at 99% in one team and 71% in another; it could be losing everything on one supplier whose statements are posted as a total; it could be recovering almost all of it late, which is a cash problem rather than a revenue one.
None of those distinctions are visible in a headline figure, and all of them have different fixes. That is why the report is worth building properly rather than estimating — an estimate can only ever produce the headline.
It also answers a question nobody asks out loud: is our leakage a decision or an accident? Firms usually assume the first, because the alternative implies a process failure. The report is the only thing that can tell them.
Build the expense half now
Upload a period's cost documents — invoices, receipts, fee notes, supplier statements — and see them come back as rows. Free tier, no registration.
Upload a document
Drop your documents here
or browse to upload
Free: 10 pages / month · Upgrade for more
Files encrypted in transit · auto-deleted after processing · GDPR compliant
The columns the report needs
| Column | Comes from | What it makes possible |
|---|---|---|
| Matter | You | Any per-matter or per-client view at all |
| Date incurred | The document | Ageing, and putting costs in the right period |
| Date document arrived | The document | Separating a slow supplier from a slow firm |
| Supplier | The document | Finding the account that leaks |
| Net / VAT / gross | The document | Reporting on whichever your treatment makes correct |
| Recoverable flag | You | Telling a waiver apart from an omission |
| Billed, and on which bill | Your billing system | The recovery half of the ratio |
| Who incurred it | You | Turning a firm number into something actionable |
Five of the eight come off the document and can be read. Three cannot — matter, flag and billing status — and no tool should pretend otherwise, because a supplier has no idea which of your matters their invoice relates to.
The two date columns are the ones most often collapsed into one, and it is a false economy. With both, a poor recovery number can be traced to a supplier who invoices six weeks late or to a firm that sat on documents for six weeks. With one, those two look identical and get the same wrong fix.
The missing denominator
Here is the failure that makes most recovery reporting worthless, and it is entirely structural.
The report divides what was recovered by what was spent. The numerator comes from the billing system, and it is complete — every bill raised is in there, because raising a bill is what the system is for.
The denominator comes from whatever costs were captured, and it is not complete. It is missing the receipt that never got submitted, the portal invoice nobody downloaded and the forty lines inside a supplier statement that was posted as one figure.
Every one of those omissions makes the ratio look better. A firm that captures 90% of its costs and recovers all of them reports 100% recovery, and is losing a tenth of its outlay. Meanwhile a firm that captures everything and recovers 92% reports a worse number while being in a considerably better position.
So the first useful thing any firm can do with this report is not to compute it. It is to compare a period's payments against a period's documents and find out how complete the denominator actually is. Until that is known, the ratio is measuring the quality of the record-keeping rather than the quality of the billing.
Reading a recovery rate honestly
| What you see | Innocent explanation | The one worth checking |
|---|---|---|
| Above 100% | Mark-up, or last period's costs billed now | Costs recharged twice |
| Suspiciously high | Genuinely tight billing | Costs missing from the denominator |
| A sudden drop | One large absorbed cost | A supplier account that stopped being split |
| Steady, slightly low | Deliberate waivers | A category nobody ever flags |
| Fine overall, poor in one team | Different work, different terms | One person's costs never reaching a matter |
| Good rate, ageing balance | Nothing wrong yet | Recovery happening too late to count |
The second row is the one that should make a finance lead uncomfortable rather than pleased. An unusually good recovery rate in a firm that has never checked its denominator is the expected result of poor cost capture, not of good billing.
The last row is the subtle one. Costs recovered eventually still show up as recovered; the report has no opinion about when. A firm with an excellent rate and a growing pile of six-month-old unbilled costs is heading somewhere the ratio will not show for another two quarters.
Four ways to cut it, in order of usefulness
By supplier. The most actionable cut and the least used. Leakage concentrates: it is usually one or two accounts, almost always the ones that send monthly statements covering many matters. Fixing how one document is processed can move the firm-wide number.
By person. Uncomfortable and informative. Recovery differences between fee earners are rarely about carelessness — they are usually about one person having a different understanding of what is recoverable, which is a policy gap wearing a personal disguise.
By cost type. Travel, couriers and printing behave differently from fees and searches. A firm that recovers court fees perfectly and travel not at all does not have a recovery problem; it has an unwritten policy on travel.
By matter or client. The cut everyone builds first and the least diagnostic, because it mixes everything together. Useful for the biggest relationships, noisy everywhere else.
Written off, or simply never billed
These produce the same accounting outcome and require opposite responses, which is why separating them is the single most valuable thing the report does.
A write-off is a decision. Somebody looked at a cost and chose not to charge it — for a good client, on a fixed fee, after a mistake of the firm's own. It is a commercial lever and firms should use it.
A cost never billed is not a decision. It is a receipt that stayed in a pocket, a statement posted as a total, a late invoice on a closed matter. Nobody weighed anything.
A firm that cannot tell them apart sees one number and reaches for the only lever it can see: tighten the policy, waive less, charge more of what is chargeable. If the actual problem was capture, that response annoys clients and changes nothing, and the money keeps leaving.
The flag that separates them costs a second per cost and is described on recoverable cost flagging. Without it, this report has one number where it needs two.
Periods, and the lag nobody adjusts for
Costs and their recovery do not happen in the same month, and a report that ignores that produces noise dressed as signal.
A cost incurred in March, invoiced by the supplier in April and billed to the client in May appears as an unrecovered cost in March, an unrecovered cost in April, and a recovery in May with no matching cost. Three months, three misleading pictures, one perfectly normal sequence.
Two things make this manageable. Report on the date incurred rather than the date the document arrived, so a cost sits in the period the work happened. And look at a rolling window — a quarter, or three months trailing — rather than a single month, because a single month is mostly lag.
The exception worth watching monthly is the ageing, not the ratio. How old is the oldest unbilled cost is a question that does not care about lag and answers a different problem.
Building it every month
1 · Read the documents
A period's invoices, receipts, fee notes and statements, up to 100 files in a pass.
2 · Split the statements
Multi-matter accounts become lines. Left as totals, they are a hole in the denominator.
3 · Allocate and flag
Matter, and recoverable / absorbed / undecided. The two things a document cannot supply.
4 · Reconcile to the bank
Payments without documents are the measure of how complete the denominator is.
5 · Join to billing
Bring in what was billed, per matter, from wherever your bills are raised.
6 · Cut it four ways
Supplier, person, cost type, matter — in that order of usefulness.
Step four is the one that makes the rest honest, and it is the one usually skipped because it is the only step that can produce bad news. The full monthly sequence, with the allocation and chasing around it, is in how to track disbursements on a matter.
Six traps in the numbers
Mixing net and gross. Recharge net, measure gross, and the report shows a permanent shortfall that is really the VAT.
Counting a supplier statement as one cost. One line in, forty costs missing, and the ratio improves.
Using the document date as the incurred date. Slow suppliers then look like a billing problem.
Averaging across teams doing different work. A litigation team and a compliance team have no business sharing a benchmark.
Treating a partial recharge as a full one. The row needs both the amount charged by the supplier and the amount recovered.
Reporting a single month. Mostly lag, occasionally alarming, rarely meaningful.
The first and the last are the most common, and both produce a number that somebody then acts on. A report that is wrong in a stable direction is more dangerous than no report, because it survives scrutiny by being consistent.
A worked example, and what fixing the denominator does to it
An illustrative practice — the figures are made up, the shape is not. It spends around £14,000 a month on client costs and has always reported recovery from what its billing system knew about.
| Line | As reported | After one honest month |
|---|---|---|
| Costs captured | £12,050 | £14,180 |
| — of which found inside supplier statements | £0 | £1,640 |
| — of which unclaimed personal spend | £0 | £310 |
| — of which portal invoices never downloaded | £0 | £180 |
| Costs billed to clients | £11,570 | £11,570 |
| Reported recovery rate | 96% | 82% |
Nothing about the practice changed between those two columns. No client paid less, no partner waived anything, no supplier put its prices up. The only difference is that the second column counted costs the first column never knew existed.
And the number got worse. This is the part that makes recovery reporting politically awkward: doing it properly for the first time always looks like a deterioration, and somebody in the room will say the old figure was fine. It was not fine; it was flattering, which is a different thing.
What the second column also produces is a target. £2,130 of cost was found that had never reached a bill. Some of it will turn out to be legitimately absorbed, and the rest is recoverable revenue that was previously invisible — worth more every month than the reporting exercise costs to run.
The ageing view, which matters more than the ratio
If a practice produces only one number from all of this, it should not be the recovery rate. It should be the age of the oldest unbilled cost.
The ratio is a backward-looking average that a good month can mask. The ageing is a live picture of what is currently at risk, and unlike the ratio it cannot be improved by leaving costs out — an omitted cost simply is not in the bucket, which is exactly what you want it not to be.
| Bucket | What it usually means | Action |
|---|---|---|
| 0–30 days | Normal billing cycle | Nothing |
| 31–60 days | Waiting for a bill, or an undecided flag | Check the flag, not the cost |
| 61–90 days | The matter is quiet or the decision was deferred | A named person decides this month |
| 90+ days | In practice, already lost | Bill it or write it off deliberately |
The last row is the one worth being blunt about. A cost that has sat unbilled for three months is almost never billed afterwards — not because it could not be, but because nobody wants to send a client an invoice for something they have forgotten agreeing to. Writing it off on purpose at least records a decision and puts the number in front of someone.
Practices that review the 61–90 bucket monthly find it stays small. Practices that only look at the ratio find the 90+ bucket grows quietly until somebody asks what the unbilled balance is made of, and nobody can answer.
Who actually reads it
The finance lead, monthly, for the ageing and the denominator completeness rather than the headline ratio.
Partners or directors, quarterly, for the split between what was waived on purpose and what was lost — which is a management question rather than an accounting one.
Whoever owns supplier relationships, because the supplier cut is where the actionable finding usually is, and it is normally one or two accounts.
Nobody, if it is a single number on a slide. A recovery percentage with no cuts underneath it gets glanced at, not used — which is the fate of most of these reports and the reason they stop being produced.
Where to start
Not with the ratio. Take one month, read every cost document into rows, and put the bank and card statements next to them. The gap between those two lists is the only number that tells you whether any recovery figure you have ever produced meant anything.
If the gap is small, compute the ratio and start cutting it by supplier. If the gap is large — and it usually is the first time — fix the capture before measuring anything, because measuring an incomplete denominator produces a number that flatters the firm in exactly the wrong direction.
The capture side is on client disbursement tracking, and how it fits a whole practice is on billing for professional services.