What this is about
A professional services bill has two components. There is the work — hours at a rate, or a fee agreed in advance. And there is everything the practice laid out along the way: court fees, searches, expert reports, travel, printing, filing fees, specialist software bought for one job, a courier at seven in the evening.
The first component is measured obsessively. Practices buy software for it, argue about it in partner meetings, and can tell you last month's figure to two decimal places by fee earner.
The second is handled by remembering. It is a shoebox, an email folder, a note in a matter file, and a fee earner three weeks later trying to recall whether the £340 expert fee was on this case or the other one with the similar name.
This page is about that second component: the process for it, the people involved, the time it takes, and what it recovers when a practice stops treating it as an afterthought. It is not about time recording, which is already solved, and it is not about trust or client account money, which is a different subject handled on bank statement conversion for law firms.
The asymmetry, and why it persists
The imbalance is not carelessness. It follows from where each half of the bill comes from.
| Time | Costs | |
|---|---|---|
| Captured by | The person doing the work | A third party you do not control |
| Captured when | The same day, usually | Whenever the supplier decides to invoice |
| Arrives as | A structured entry | A PDF, a portal download or a photograph |
| Already knows the matter | Yes — you typed it | No, and it never will |
| Chased by | A weekly timesheet report | Nobody in particular |
| Missing entries are | Visible as a gap | Invisible — nothing shows an absence |
The last row is the whole problem. An unrecorded hour shows up as a thin day on a timesheet report, and somebody asks about it. An unrecorded cost produces no evidence whatsoever — the receipt simply stays in a coat pocket, and the practice's records are internally consistent and quietly wrong.
This is why the fix cannot be exhortation. Telling people to be diligent about receipts has been tried by every practice that has ever existed, and it fails for the same reason each time: the failure is silent, so there is nothing to be diligent in response to.
A month of cost documents in one pass
Five statements dropped in together, read, merged into a single table and exported to Excel — the same shape as a practice's monthly cost run.
Up to 100 files per export
Every row keeps its source file and page
Excel, CSV, JSON or XML
Who does what
The single most common reason a cost process fails is that nobody owns it. Four roles, and only one of them is new.
The fee earner
Incurs the cost and knows the matter. Their only job is to make the document reach the practice with the matter attached — nothing else about the process should be theirs.
The cashier or bookkeeper
Runs the monthly read, splits the supplier statements, and produces the allocated table. This is where the work actually lives.
The matter partner
Decides the borderline calls: what gets recharged, what gets absorbed for a good client, what was the practice's own mistake.
The finance lead
Owns the recovery number and the ageing, and is the only person who ever looks at the whole picture rather than one matter.
The split that matters is between the second and third. Reading and allocating is clerical and should be done by whoever is fastest at it. Deciding what a client pays for is a commercial judgement and should never be made by whoever happens to be holding the receipt.
Practices that collapse those two into one role get one of two outcomes: a cashier quietly making commercial decisions, or a partner spending a morning on data entry. Both are expensive in different directions.
The monthly cycle
A workable rhythm for a practice of ten to forty fee earners. Bigger practices run the same sequence weekly; smaller ones run it monthly and it still works.
Throughout
Documents arrive
Into one place — a shared mailbox, a folder, an upload. The rule is that a cost document is never stored in a person, only in the place.
Day 1
The read
Everything gathered for the period goes through in one pass and comes back as rows: date, supplier, net, VAT, gross, source file and page.
Day 1
Statements split
Any supplier account covering several matters becomes lines. Left as a total, those costs never reach a client and never come back.
Day 1–2
Allocation
Matter attached to every row. Whatever cannot be placed goes on a short unallocated list rather than being guessed at.
Day 2
The unallocated list circulates
Ten or twenty rows to the fee earners, answered in an afternoon. This is the only part that touches people doing billable work.
Day 2
Flagging
Recoverable, absorbed, or not yet decided. The third state is what keeps undecided costs from silently becoming waivers.
Day 3
Reconcile to the bank
Payments with no document behind them are the measure of what the process is still missing. This step is the one that produces bad news, and skipping it is why most processes drift.
Day 3
Into billing
Allocated, flagged, recoverable rows go to whatever raises the bills, as lines rather than as a lump.
Day 4
Ageing review
The oldest unbilled costs, looked at as a list. Anything past sixty days gets a decision rather than another month of waiting.
Four days elapsed, but only about a day of actual work, and most of it on days one and two. The step-by-step version, with the chasing and the edge cases, is in how to track disbursements on a matter.
Where the hours actually go
A practice with around 300 cost documents a month, before and after. These are the proportions we see described rather than a promise about your practice.
| Activity | Manual | With the read automated |
|---|---|---|
| Typing documents into a sheet | 9–12 hrs | ~20 min |
| Splitting supplier statements | 3–4 hrs | ~15 min |
| Allocating to matters | 3–4 hrs | 3–4 hrs |
| Chasing missing documents | 2–3 hrs | 1–2 hrs |
| Checking totals and typos | 2 hrs | ~30 min |
| Bank reconciliation of costs | 2 hrs | 1 hr |
The row that does not move is the important one. Allocation stays exactly where it was, because attaching a cost to a matter is a judgement about your practice that no document contains and no software can infer. Anyone claiming otherwise is describing a guess with a confidence score attached.
What changes is the ratio. A process that was three-quarters typing and one-quarter judgement becomes almost entirely judgement — which is both cheaper and, for the person doing it, considerably less demoralising.
What it actually recovers
Two separate returns, and practices usually notice the smaller one first.
The time saved
Around fifteen hours a month at the volumes above. Real, easy to calculate, and the one that gets into the business case — but it is the less interesting half, because the hours saved belong to a cashier rather than a fee earner.
The costs that now reach a bill
This is the larger number and the harder one to face, because measuring it means admitting what was being lost. A practice spending £14,000 a month on client costs and capturing 90% of them is giving away £1,400 a month — not as a discount, not as a commercial gesture, but by accident.
That figure is pure margin. It was already paid out; recovering it costs nothing further. Ten hours of a cashier's time saved is worth a few hundred pounds; the recovered costs are worth several times that and compound every month the process holds.
The honest version of the calculation
You cannot compute the second number in advance, because the whole point is that the missing costs are invisible. What you can do is run one month properly — every document read, statements split, bank reconciled — and compare the total against what the same month would normally have produced. The gap is the answer, and it is the only trustworthy version of it.
How to build that comparison, and why the ratio flatters practices with poor capture, is on the time and expense recovery report.
Four practices, four different problems
A law firm
High volume of small third-party fees — searches, filing fees, court fees — arriving from a handful of suppliers as monthly statements covering dozens of matters. The whole battle is fought inside those statements: split them into lines and the recovery problem largely disappears; post them as totals and no amount of diligence elsewhere helps.
An architectural practice
Fewer, larger costs — models, printing, planning fees, specialist consultants — but spread across projects that run for years. The difficulty is not volume, it is the lag: a cost incurred in month three of a thirty-month project has to survive until the stage it belongs to is billed, and things that must survive that long need to be written down rather than remembered.
A consultancy
Travel dominates, and travel is the category most often absorbed by accident because it arrives as photographed receipts from people who are somewhere else. Consultancies also renegotiate expense terms per engagement more than other practices, which means the recoverable flag genuinely varies client by client and cannot be a firm-wide default.
A surveying or engineering practice
A mixture: recurring small costs, occasional large specialist fees, and a high proportion of work under fixed fees where the recharge question is answered by the contract rather than by habit. These practices need the flag more than anyone, because the same cost type is recoverable on one engagement and absorbed on the next.
The common thread is that none of these are volume problems in the way an invoice processing backlog is a volume problem. They are all continuity problems: a fact known at one moment by one person has to survive intact until a bill is raised weeks or years later.
Fixed fees change the question, not the need
A common objection: we work on fixed fees, so recharges do not apply and none of this matters.
The first half is often true and the conclusion does not follow. Under a fixed fee the costs are not recovered separately — they come out of the fee — which means they are not a billing input but a margin input. They now determine whether the engagement made money.
A practice that does not capture costs under a fixed fee has no idea which of its fixed fees were priced correctly. It knows revenue and it knows time; the third term is missing, so every profitability conversation is conducted with one number absent.
This is the situation where the tracking changes from a recovery exercise into a pricing one. The costs still need capturing, allocating and totalling per engagement — they simply flow into next year's fee rather than this month's bill.
It is also the case where the three-state flag earns its keep, because “absorbed under a fixed fee” is a completely different fact from “nobody has decided yet”, and they look identical in any system that only offers a tick box.
Everything traces back to the engagement letter
Every argument about whether a cost is recoverable is really an argument about what was agreed at the start, and most practices discover their engagement terms are vaguer than they remembered at precisely the wrong moment.
Costs recharged at cost, or with a handling uplift?
Decides whether the row needs one amount or two.
Net or gross, given the client's VAT position?
Decides which column is the recharge.
Is there a threshold above which the client is asked first?
Decides whether an unapproved cost is recoverable at all.
Travel — all of it, or only beyond a stated radius?
The single most disputed category, in every profession.
Internal charges: printing, copying, storage?
Increasingly hard to defend without an explicit term.
Practices that answer these five once, in the engagement template, stop having the same conversation on every matter. Practices that do not end up with a recoverable flag that means “whatever the person setting it believed on the day”, which is not a flag at all.
The bill that gets queried
A client asks what a £480 charge on their bill was for. This moment is the real test of a cost process, and it is where the difference between a good one and a bad one becomes visible to the client.
In a practice with a proper record, the answer takes under a minute: the date, the supplier, the amount, the matter, and the source document the row came from. It is a factual reply and the query ends.
In a practice without one, someone searches an email folder, then a shared drive, then asks the fee earner, who is in a meeting. Two days later a partial answer goes back. Sometimes the charge simply gets credited to make the conversation stop — which converts a legitimate cost into a write-off through nothing but poor filing.
That is why keeping the source file and page reference on every row matters more than it sounds. It is not an audit nicety; it is the thing that turns a query into a one-minute reply instead of a credit note.
How costs should appear on the bill
Capturing costs correctly and presenting them badly wastes most of the benefit, because a bill that invites questions gets paid late whatever the underlying record looks like.
There are three usual presentations and they behave very differently.
| Presentation | Effect |
|---|---|
| One line, labelled Disbursements | Fastest to produce, most likely to be queried, and the hardest to defend when it is |
| Grouped by type, with a total per group | Usually the right balance — enough detail to look considered, short enough to read |
| Every cost itemised on the face of the bill | Complete, and on a long matter it buries the fee under forty lines of couriers |
The pattern that works in most practices is the middle one on the bill, with the full itemised list available on request. That combination answers the two questions a client actually has — what was this for, and can you show me — without making them read a ledger to approve an invoice.
It also has a quiet second effect. Producing grouped totals requires the costs to have a type on them, which means somebody categorised them, which means the practice ends up with the very field it needs to find out where its recovery is leaking. The billing requirement enforces the reporting discipline.
One thing worth avoiding: a bill that shows costs net when the client expects gross, or the reverse. It is a formatting decision rather than a commercial one, and it generates more queries than any amount ever charged.
Three numbers a practice should be able to state
Not a dashboard. Three figures that a managing partner should be able to give without notice, and almost none can.
What we spent on clients last month
Total outlay on clients' behalf. Most practices can only give what was reimbursed or coded, which is a different and smaller number.
How much of it has been billed
The recovery figure — but only meaningful once the first number is complete, which is why it comes second rather than first.
How old the oldest unbilled cost is
The one that predicts next quarter. It cannot be improved by leaving costs out, which makes it the most honest of the three.
The order matters and it is the reverse of how these are usually built. Practices start with the recovery percentage because it sounds like the answer, then discover it rests on a base nobody has verified. The first number is dull and it is the foundation; the third is the early warning; the second is only trustworthy once the other two exist.
A practice that can state all three has, by definition, a working cost process — there is no way to produce them from memory. Which makes them a better test of the process than any audit of it.
Four objections worth taking seriously
Our practice management system already does this
It has fields for it, which is not the same thing. The fields are there and empty, because filling them means somebody reading a PDF and typing. The gap is not in the destination, it is in getting data to the destination.
Our volumes are too low to bother
Below roughly a hundred cost documents a month, that is probably right and we would rather say so than sell you something. The threshold is not the typing, it is the point where a supplier statement covering many matters becomes routine.
Fee earners will not change how they work
They should not have to. The only behaviour asked of them is sending a document to one place with a matter reference — which is less than any receipt policy already demands. Everything downstream sits with the cashier.
We tried a shared spreadsheet and it lapsed
Spreadsheets lapse because the entry is manual and the failure is silent — the two conditions that guarantee decay. Removing the typing removes the first; the monthly bank reconciliation removes the second by making omissions visible.
Starting on one matter
Do not roll this out across a practice. Take one active matter with a reasonable amount of third-party cost — a piece of litigation, a project in construction, an engagement with real travel — and run the full cycle on it for one month.
Read every cost document for that matter. Split any supplier statement that touches it. Allocate and flag every row. Then put the total next to what the matter would have been billed for costs under the normal process.
That single comparison settles the question better than any argument, and it takes an afternoon. If the two numbers match, the practice's informal process is working and you can stop reading. In our experience they do not match, and the direction is never the flattering one.
From there the sequence is: the same month across all matters, then the bank reconciliation that tells you how complete the capture really is, then the recovery report once the denominator can be trusted.