Offboarding is a workflow, not a conversation. Nine steps close a client properly: bill the work in progress before you do anything else, issue a disengagement letter setting out what you have and have not done, respond to the professional clearance request within a stated service level, hand over records to an agreed list, remove HMRC agent authorisations service by service, reassign or cancel software licences, close open jobs and recurring tasks in practice management, mark the anti-money-laundering file with the relationship end date because the five-year retention clock starts there, and record the reason for leaving. On a 420-client firm losing 25 clients a year, doing this badly costs roughly £14,000 a year in written-off work in progress, unpaid final fees and unbilled admin time.
Firms have an onboarding process and an exit anecdote
Ask a practice owner how a new client is taken on and you will usually get a process: the engagement letter, the identification checks, the authorisations, the software set-up, the first-90-days plan. Ask how a departing client is closed down and you get a shrug and a story about one that went badly.
That asymmetry is not about caring less. It is that onboarding has an obvious owner and an obvious payoff, and offboarding has neither. Nobody is measured on it, nobody has written it down, and it happens at the precise moment everyone involved would rather think about something else.
The result is that a firm losing five per cent of its book a year runs the same messy, expensive scramble twenty times, reinvents it each time, and never sees the cost because the cost is distributed across half a dozen people in ten-minute pieces.
What a scruffy exit actually costs
Take a firm with 420 clients and annual fees of £1.4m, losing 6% of its clients a year. That is 25 exits. Illustrative figures, but the shape is the one I see most often.
| Leak | Per exit | × 25 |
|---|---|---|
| Work in progress written off because nobody billed before the handover | £310 | £7,750 |
| Unbilled admin: clearance, records, authorisations, chasing (3 hours at £48) | £144 | £3,600 |
| Final fees never collected (4 of the 25, at £640) | — | £2,560 |
| Annual cost | £13,910 |
Just under £14,000, which is roughly 1% of the fee line, and none of it appears as a line on any report. It shows up as write-offs, as recoverability, and as the general sense that admin time has crept up again.
The work in progress line is the one worth staring at. It is the largest, and it is entirely a sequencing problem: bill the work before the records leave, and it disappears. That is the same discipline that drives lock-up, and firms that have fixed one usually have not thought to apply it to the other.
The nine steps
This is the whole process. It should live in your practice management system as a job template with nine tasks, triggered the moment a departure is confirmed, and it should take one person about fifteen minutes plus whatever the records handover genuinely requires.
- Bill the work in progress. First, before anything else. Once records have gone and clearance has been answered, the leverage and the goodwill have both gone with them. Raise the final invoice on the day the departure is confirmed, and be specific about what it covers.
- Issue a disengagement letter. Not a courtesy note. It states the date the engagement ends, what work has been completed, what has explicitly not been done, what deadlines are outstanding and whose responsibility they now are, and what you will and will not do afterwards. This is the document that answers the question three years later about who was supposed to file the return.
- Answer the professional clearance request to a service level. Pick a number — five working days is a reasonable one — and hold to it. Firms are judged in their local market by how they behave on the way out far more than on the way in, and the incoming accountant is a referrer you have not met yet.
- Hand over records against a written list. Last accounts and the computations behind them, the trial balance and closing balances, capital allowances and fixed asset schedules, payroll and pension records, VAT history, and software access. Send the list with the records so both sides can see what arrived.
- Remove HMRC agent authorisations, service by service. Corporation Tax, Self Assessment, VAT, PAYE and CIS are separate authorisations, and removing one does not remove the others. Authorisations left in place mean you keep receiving correspondence and notices for a client who is no longer yours, which is both a nuisance and a data protection problem.
- Reassign or release software licences. Where the subscription sits with your firm, decide with the client whether it transfers or ends, and get it off your billing. Unclaimed seats are the quietest recurring cost in a practice, because nobody ever reviews a subscription that has not changed.
- Close the jobs and stop the recurring tasks. Open jobs on a departed client distort your capacity picture, and a recurring task that keeps generating is how a firm ends up with a job planner nobody trusts.
- Mark the anti-money-laundering file with the relationship end date. Under the Money Laundering Regulations 2017 the record-keeping clock runs for five years from the date you know, or have reasonable grounds to believe, the business relationship has ended — and once the period expires you must delete the personal data unless another legal reason requires you to keep it. If the end date is not recorded, neither the retention nor the deletion can be run correctly.
- Record the reason for leaving, in one field. Price, service, sold the business, retired, moved to a specialist, in-housed it, no reason given. One dropdown. This is the step everyone skips and the only one that changes anything next year.
The reason field is the whole point
Eight of those steps stop money leaking. The ninth is the only one that tells you why the leak is there.
Twenty-five departures with a recorded reason is a report. Read it once a quarter and it answers questions that are otherwise pure opinion: whether your price rise landed or cost you a cohort, whether departures cluster in one manager’s portfolio, whether the clients you lose are the ones you would have chosen to lose, and whether a service gap is sending work to a specialist you could have partnered with.
Without it, churn is weather. With it, churn is data, and it is the cheapest client research a firm will ever do — because these are people who have just decided something about you and have no reason left to be polite about it.
Deciding who to let go
Not every departure should be resisted. Most firms carry a tail of clients who cost more to serve than they pay, and the honest test is contribution per client rather than fee per client: the fee, less the time actually recorded, less the unrecorded time nobody logs.
Where a planned exit is the right answer, the sequence matters. Reprice first and let them choose — a client who accepts a proper fee is a different client, and one who declines has made the decision for you without a difficult conversation. Give notice that clears the next filing deadline rather than landing on top of it. Offer two names of firms that suit them better. And run the same nine steps, because a client you asked to leave deserves a tidier exit than one who left of their own accord, not a worse one.
The connection to profitability is direct, and it is the subject of why busy firms are not always profitable: capacity released from a loss-making client is only worth something if you have decided in advance what it is for.
Building it once
This is a small piece of work with an unusually good return. One afternoon produces a job template with nine tasks, a disengagement letter template, a records handover checklist, and a reason-for-leaving field on the client record. After that it runs itself, and the person doing it does not have to remember anything.
It also belongs next to its twin. A firm with a documented onboarding process and no offboarding process has built half a client lifecycle, and the missing half is the one that carries the money and the risk. If consistent onboarding is already in place, this is the same exercise in reverse and will take a fraction of the time. If neither exists, building SOPs is where to start, and offboarding is a good first one precisely because it is short, self-contained and currently costing you money.
Then measure one thing: the number of days between a departure being confirmed and the file being fully closed. If that number is under ten, the process is working. If nobody can tell you what it is, that is the answer.
Common questions
How long should we take to respond to a professional clearance letter?
Set a service level and publish it internally — five working days is achievable for a normal file and is well inside what the incoming firm expects. The reason to be quick is not politeness. Every clearance request comes from another practice in your market who is forming a view of how you operate, and a firm that answers promptly and completely is remembered as one worth referring to. A firm that takes six weeks and sends a partial pack is remembered differently. The only legitimate reason to delay is an unsettled account, and even then the right move is to say so immediately rather than to go quiet.
Can we hold records back if the client owes us money?
There is a legal concept of a lien over certain documents, but treat it as a last resort and take your own advice before relying on it. The practical position is that it applies only to some categories of document, it does not extend to records the client is legally required to have, and it very rarely produces payment — what it usually produces is a complaint to your professional body and a story that circulates locally. The far better control is upstream: bill the work in progress on the day the departure is confirmed, before the handover starts. Almost every fee dispute at exit is created by doing those two things in the wrong order.
Do we have to delete the client's data when they leave?
Not immediately, and not all of it. Under the Money Laundering Regulations 2017 the customer due diligence records must be kept for five years beginning on the date you know, or have reasonable grounds to believe, the business relationship has ended, and once that period expires you must delete the personal data unless you are required to retain it for another legal reason. Other records have their own retention periods. What matters operationally is that the end date is recorded on the file, because every retention and deletion decision afterwards is measured from it, and a file with no end date cannot be managed correctly in either direction.
Should we ask departing clients why they left?
Yes, but keep it to one field on the client record rather than an exit survey nobody completes. A single dropdown — price, service, sold, retired, in-housed, moved to a specialist, no reason given — filled in by whoever handles the exit takes ten seconds and produces a quarterly report that is genuinely decision-grade. Where a departure is significant, a short call from someone other than the client's usual manager will often get a straighter answer than the reason first given, because people soften the truth for the person they have been dealing with. Track the pattern rather than agonising over individual cases.
Let’s build a firm that runs without you in the middle of it.
A confidential, no-obligation call to understand your firm, where it’s stuck, and whether Optivo is the right fit. If it’s not, I’ll tell you.