Expense management for small teams has a narrow path to walk. Too little control and you find out what was spent after it is gone. Too much and people stop buying things they genuinely need, or route around the process entirely.
The reimbursement problem
Most small teams start with reimbursement: staff spend their own money and claim it back. It requires no setup, and it has three costs that show up later.
- Staff are financing the business, which is unfair and quietly resented.
- Claims arrive in batches, so you learn about spending weeks after it happened.
- Receipts go missing, and the record depends on someone's memory.
Cards with limits, not blanket trust
Issuing cards with per-person limits changes the shape of the problem. Spending is visible as it happens rather than reported afterwards, and the limit does the controlling instead of a policy nobody reads.
Useful limits are specific.
- A monthly ceiling per person, set to their actual role.
- Category restrictions where they genuinely apply.
- A single-transaction limit above which approval is required.
A limit is a decision made once, in advance, calmly. An approval request is a decision made repeatedly, under time pressure, by someone who is busy.
Who should hold a card
The instinct is to issue cards narrowly and make everyone else claim. That usually costs more than it saves, because the people spending most often are the ones sent through the slowest route.
A more useful test is whether someone spends on the company's behalf regularly and predictably.
- Anyone buying stock or supplies, however junior.
- Anyone who travels for work more than occasionally.
- Branch or site managers, who otherwise become a bottleneck for their whole team.
Set the limit to match the role rather than to signal trust. A modest limit on many cards is safer than a large limit on a few, because exposure is contained and visibility is complete.
Subscriptions, the quiet leak
Recurring software charges are the spending category that grows without anyone deciding to grow it. Tools get trialled, adopted by one person, and continue billing long after they stopped being used.
Review recurring charges quarterly and ask a single question about each one: who used this last month. Anything without an answer should be cancelled rather than renewed by default.
Categories that mean something
Keep the category list short enough that people classify correctly without thinking. Ten categories used properly beat forty used carelessly.
Name them after what the business does rather than after accounting conventions. Someone spending money at a fuel station should not have to work out which ledger it belongs to.
Receipts, at the moment of spending
The only reliable time to capture a receipt is immediately. Any process that involves keeping paper until later produces gaps, and the gaps are always in the months you most need the record.
The approval trail
Every approval should record who gave it and when, automatically. Not because you expect disputes, but because the one time it matters, reconstructing it is impossible.
This is also what makes expense management for small teams survive an audit. The trail exists whether or not anyone asks for it.
A spend policy people will actually read
Long policies are not read, and unread policies are not followed. One page is enough for most small teams, and it should answer four questions.
- What can I buy without asking?
- What needs approval, and from whom?
- What is never reimbursable?
- What do I have to keep, and by when?
Anything that does not answer one of those is background material, not policy. If people can hold the rules in their head, they will follow them.
VAT on expenses
Expense records that ignore VAT create work at exactly the wrong moment. Capture the VAT element at the point of spending rather than reconstructing it later from receipts that may no longer exist.
This is a small habit with a large payoff, because reconstructing VAT across a quarter of small purchases is one of the least rewarding tasks in a finance team's month.
What to watch for
Controls exist partly to catch honest error and partly to make dishonesty difficult. A few patterns deserve a second look, none of which are accusations on their own.
- Spending that repeatedly lands just below an approval threshold.
- The same amount recurring to the same merchant with no matching business need.
- Receipts submitted well after the spending date, consistently.
- One person able to both create a supplier and pay it.
That last one is a structural issue rather than a behavioural one, and it is the easiest of the four to fix.
Closing the month
Expenses are usually the reason month end drags, because the information arrives last. Two changes fix most of it.
- A cut-off date after which spending belongs to the following month, applied without exception.
- Categorisation at the point of spending, so nothing is sitting uncategorised when you want to close.
Connecting spend to the books
Expenses that live in a separate tool have to be re-entered somewhere, and every re-entry is a chance to be wrong. When spending, categorising and recording happen in one place, month end has nothing to reconstruct.
See expenses, expense cards and spend controls, or the expense management software overview.
