Outsourcing the finance function is usually adopted for one of two reasons: the business has grown past what one bookkeeper can hold together, or it has never built a finance team at all and does not want to. Both are reasonable starting points. Where outsourcing goes wrong is not usually in the decision to do it — it is in treating the arrangement as low-cost bookkeeping rather than as a function with its own boundaries, its own handover requirements, and its own points that management should never actually give away.
What can genuinely be outsourced
The transaction-processing and reporting layer of a finance function is the part that outsources cleanly, because it is procedural rather than judgemental: recording transactions as they occur, reconciling bank, supplier and customer accounts on an agreed cycle, running payroll from an agreed input with the deductions and statements that follow from it, closing the books to a defined date every month, and producing management accounts and the reconciliations that stand behind them. Periodic returns prepared from a ledger that has actually been closed — rather than from a summary reconstructed separately — belong here too, because the return and the books cannot then disagree with each other.
Cover through a change of finance staff, or through a change of accounting system, is also a legitimate reason to outsource, even where a business otherwise runs its own finance team: an external provider can hold the ledger together through a transition that would otherwise stall on a single person's availability.
What should usually stay under management's own control
Three things do not belong to the outsourced function, whatever else is delegated to it. Approval authority — who can authorise a payment, sign off an expense above a limit, or approve a journal — is a management decision, not an operational one, and it should not migrate to whoever happens to be processing the transaction. Banking access and mandates — who can move money, and under what limits — sit with the business for the same reason. Strategic and structural decisions — a new entity, a change in ownership, a financing arrangement — are informed by the numbers an outsourced function produces, but the decision itself is management's, not the provider's.
The practical test is straightforward: an outsourced provider should be able to tell you precisely what the numbers say and reconcile them to the underlying records. It should not be the one deciding, unsupervised, what gets paid, approved or committed.
Transaction processing versus management reporting
These are two different outputs, produced from the same ledger, and a business should be clear about which it is actually receiving. Transaction processing is the recording layer — every entry posted, every reconciliation performed, on schedule. Management reporting is what a business actually acts on: management accounts that arrive on a fixed date every month, not whenever the underlying processing happens to be finished, with enough detail to support a decision rather than merely to confirm that the books balance.
A finance function that only processes transactions well, without ever producing management reporting a business can act on, has not actually solved the problem that usually motivates outsourcing in the first place. "We already send the numbers" and "management gets what it needs to act by a fixed date" are not the same claim, and only the second is worth measuring an arrangement against.
Reconciliations and receivables/payables support
Bank, supplier and customer reconciliations performed on schedule — by someone who did not also make the original entries — are where most of the value in an outsourced bookkeeping function actually sits, because they are what turns a ledger into a set of figures management can rely on. Receivables and payables support that is properly part of this work means ageing that is current, credit terms that are actually being enforced against what is recorded rather than only against what is remembered, and returns or credit notes that go through an approval step rather than being raised on request.
Month-end close and management accounts
A close with a fixed date, every month, is the deliverable that most outsourcing arrangements are actually bought to produce, and it is also the one most often delivered inconsistently. A close that arrives "when it arrives" describes a month that has already ended by the time anyone can act on what it says. The management accounts that come out of the close, and the reconciliations that stand behind each figure in them, are what should be reviewed before an arrangement is judged to be working — not simply whether a file was sent.
Payroll, and where compliance coordination actually fits
Payroll processed from an agreed input, with deductions and the statements that follow correctly produced from the same run, is squarely within what an outsourced finance function does. Compliance work — periodic returns — belongs here too, but specifically as work prepared from the closed ledger, not as a separate tax-advisory function bolted on to bookkeeping. Where a position is genuinely uncertain, or a notice has arrived, that is advisory or audit-response work in its own right, coordinated with the finance function rather than absorbed into it.
Segregation of duties, when the whole ledger sits with one provider
Handing the ledger to an outside provider does not remove the need for segregation of duties — it relocates the question of where the segregation actually sits. Whoever posts an entry should not be the same person reconciling it. Whoever runs payroll should not be the same person approving the input that determines what gets paid. A business that outsources the entire bookkeeping function to a single individual, inside or outside the business, has simply moved the single-point-of-failure risk it was trying to solve, not removed it. This is worth asking about explicitly when the arrangement is set up, not assumed from the fact that the work is now "outsourced" and therefore presumed to be controlled.
Access to banking, ERP and data
What a provider can see and do inside the business's own systems should be decided deliberately, not left to whatever access happened to be granted when the relationship started. Read access to the ledger for processing and reconciliation is one thing; payment initiation rights are another; and the two should not be granted to the same access profile by default. The same applies to the underlying data itself — a business should know, from the outset, where its accounting data actually resides while an external provider is maintaining it, and what happens to that data and its accessibility if the arrangement ends.
Documentation and handover
An outsourced finance function that exists only in one provider's working files is a dependency a business does not notice until the arrangement ends and the records have to move. The chart of accounts, the reconciliation methodology, the payroll calculation basis and the cut-off conventions used at each close should be documented well enough that a different provider, or an internal team, could pick up the ledger without reconstructing it from scratch. This is worth confirming at the start of an engagement, not requesting for the first time when it is ending.
How the arrangement should change as the business grows
An outsourcing arrangement sized for a business with one legal entity and a simple product range does not automatically scale to one running several entities, a second currency, or a materially larger transaction volume. The signs that the arrangement needs to be reviewed rather than simply continued are usually the same ones that would justify building an internal function in the first place: management accounts that take longer to reconcile each month rather than shorter, a close that increasingly depends on manual workarounds outside the system, or decisions that are being made without waiting for the numbers because the numbers no longer arrive quickly enough to inform them.
Bookkeeping and audit stay separate
Where a practice maintains a company's accounting records, it should not also audit that company's financial statements — reviewing one's own work is not a review. Which of the two is wanted is a decision to settle at the start of a relationship, not at a year end when the two roles would otherwise collide.
What management should do
- Decide, in writing, what stays under management control — approval authority, banking mandates and strategic decisions — before the arrangement starts, not after a problem exposes that nobody decided.
- Judge the arrangement on management reporting, not on transaction processing alone. A ledger that balances is not the same claim as management accounts a business can act on by a fixed date.
- Confirm segregation of duties explicitly, inside the provider's own team, rather than assuming that "outsourced" implies "controlled."
- Set access levels deliberately — processing and reconciliation access is not the same as payment initiation access, and the two should not default to the same profile.
- Require documentation of the chart of accounts, close conventions and payroll basis from the start of the engagement, not as an afterthought when the relationship ends.
- Review the arrangement's fit periodically as the business grows, rather than assuming a structure sized for an earlier stage of the business still fits.
This is the shape our own outsourced finance work takes — routine processing, a close with a date on it, and a clear line kept between bookkeeping and any audit work, so neither is mistaken for covering the other. If a close keeps slipping, or a finance function needs to be handed over cleanly, tell us roughly what the business runs today and where the strain actually is, and we can set out what a properly bounded arrangement would look like.
This note is general information, not advice on your particular circumstances. Tax law and deadlines change — please confirm the position before acting on it.