Why agency expenses go missing from project costs

An agency can have every transaction recorded correctly and still not know what a client project actually cost. A card charge lands in the accounting system, a receipt sits in someone’s inbox, and the project code gets added later, if anyone remembers. By month-end, the finance lead or account manager is trying to connect those pieces from memory.

This gets harder when one purchase serves more than one account. A stock asset may support several campaigns. A freelancer may work across multiple client projects in the same week. A travel expense might cover a client meeting and an internal planning day. If the expense is coded to a general category without a project allocation, the books may balance while project margins stay wrong.

For a marketing and creative agency, that gap affects more than bookkeeping. It changes what an owner sees when deciding whether to renew a client, adjust a retainer, scope the next campaign, or change how the team delivers work. An expense that should have been charged to a client can quietly become an agency cost. An expense assigned to the wrong project can make a profitable account look unprofitable, or the reverse.

Agencies may have leakage to investigate. Treat that as a range to investigate, not a promise that expense allocation alone will recover the full amount. Some leakage comes from missed billables, weak project scoping, or delivery overruns. Better expense matching gives you a more dependable view of where the money is going and where a follow-up is needed.

What the manual process looks like

In many agencies, the work is split across card statements, receipt capture, accounting software, project management tools, and spreadsheets. Each tool may hold a piece of the story, but no one has time to reconcile every transaction against project activity as it happens.

A typical month-end cleanup might involve:

Asking a cardholder to identify a charge from a vendor name that doesn’t match the receipt.
Searching email or a shared drive for a receipt that was never attached to the transaction.
Comparing the purchase date with project schedules, campaign briefs, or a freelancer’s timesheet.
Asking an account manager whether a subscription, asset, meal, or travel cost belongs to a client.
Splitting one transaction across several projects or marking it as internal.
Updating the accounting record, then separately updating a project budget or margin report.

The delay matters. An account manager answering a question three weeks later may not remember whether a $240 asset purchase was for one client, several clients, or a pitch. The finance team may then choose a generic overhead code to close the month. That saves time in the moment, but it hides the cost from the project report.

Expense allocation also connects to the work of automating agency project profitability. If direct costs and labor aren’t assigned consistently, the margin calculation is only as good as the inputs. Fixing the expense trail is one practical step toward a more useful account-level view.

How an expense-allocation agent works end to end

An expense-allocation agent doesn’t need to make every decision on its own. Its job is to gather the relevant evidence, suggest a project or internal code, and route uncertainty to the person who can resolve it. A good setup keeps a human in control of ambiguous or material expenses while removing repetitive lookup and data entry.

1. Bring transactions and receipts together

The workflow starts with card transactions and receipt records from the agency’s existing systems. Each transaction is matched against receipt details such as merchant, date, amount, tax, and currency. When the receipt is missing, the workflow can flag the gap and request it from the cardholder.

Matching should allow for ordinary differences. A receipt may show a shortened vendor name, the transaction may post a day later, or the final charge may differ slightly because of tax or currency conversion. The agent can use those signals to suggest a match, while keeping the original transaction and receipt available for review.

For agencies that still enter receipt details by hand, automating receipt entry at a marketing agency can address that upstream step. Entry and allocation solve different parts of the problem. One captures what was purchased, the other determines which project should carry the cost.

2. Suggest the right client or internal project

Once a receipt is matched, the agent looks for project clues. These might include a project code on the receipt, a client or campaign name in the purchase description, the cardholder’s assigned accounts, the work scheduled around the purchase date, or an existing vendor-to-project pattern.

For example, a designer makes a stock-image purchase during a week when they’re working on two client campaigns. The vendor and cardholder alone aren’t enough to assign the cost. The agent can compare the receipt note, project calendar, and any campaign reference in the transaction details. If one project is a strong match, it can recommend that code. If both projects are plausible, it should ask for a split or route the expense to review.

Internal work needs the same care. A charge might belong to agency marketing, a new-business pitch, staff training, or general overhead. Leaving these expenses in an unreviewed catch-all makes it harder to compare delivery costs with the cost of winning and running the business.

3. Apply allocation rules and flag exceptions

Agencies can define rules for costs that recur. A monthly tool subscription may be assigned to a specific internal function, while a client-specific production cost may follow the project code on the brief. Shared expenses can be split using a rule the agency has approved, such as a documented allocation percentage or a named set of projects.

Rules should not turn guesses into facts. The workflow should flag an expense when the receipt is missing, the amount is outside an expected range, a new vendor appears, a client or project can’t be identified, or the proposed split has no supporting instruction. A human can then approve, correct, or reject the suggestion. The correction becomes a useful reference for future transactions, subject to the agency’s review process.

Keep an audit trail of the original transaction, receipt, suggested allocation, approval, and any change made. That makes it easier to explain the code later, and it helps prevent a corrected expense from quietly reverting to its earlier category.

4. Update records and publish a review queue

After approval, the allocation can be sent to the relevant accounting or project record, depending on the systems and permissions in place. The workflow should not silently overwrite a closed period or change a client-billable setting without approval. Those are policy decisions, not routine data cleanup.

A short exception queue is more useful than a long list of every transaction. Show the amount, cardholder, receipt status, likely project, reason for review, and next action. A finance lead can resolve a missing receipt, while the account manager can confirm whether a cost belongs to a particular client. Keep routine, high-confidence matches out of that queue so the team spends its time on the decisions that need judgment.

For another related process, automating client expense reimbursement shows how a clear approval trail can help when costs need to be submitted or paid back. The same principle applies inside an agency: make ownership and evidence visible before a transaction reaches month-end.

A practical example

Say a paid media specialist uses a company card to buy a $680 research report. The transaction arrives with a shortened vendor name, and the receipt is attached two days later. The agency has three active client campaigns that could use the report, plus an internal market review.

A manual process might leave the charge uncoded until the account manager sees it on a month-end spreadsheet. An agent can match the receipt to the transaction, identify the cardholder’s active projects, and check the receipt note or purchase request for a client reference. If it finds a clear match, it proposes the project and records the evidence. If the clues point to two campaigns, it asks the owner to confirm whether to split the charge or treat it as an internal cost.

The important result isn’t that the system makes every call automatically. It’s that the cost is surfaced while the context is still available, and the person making the decision sees the evidence rather than starting from a blank row.

Book an Omni Audit to map where this workflow would fit across your transaction, receipt, and project records.

What to measure after the change

Before automating, record a baseline for a normal month. You don’t need a complicated scorecard. A few operational measures will show whether the process is improving:

How many transactions are missing a project or internal code when the month closes?
How much staff time goes into chasing receipts and asking what a purchase was for?
How many expenses need recoding after project reports have been shared?
How much time passes between a transaction posting and its allocation being approved?
How often do project costs remain in a general overhead category despite having a likely owner?

Compare the same measures after the workflow has been in use long enough to cover a full billing cycle. Don’t count every proposed allocation as a saving. Look for fewer unresolved items, less rework, and project reports that reflect the costs your teams actually incurred.

This also gives owners a clearer conversation with account managers. Instead of asking why a project margin changed after the fact, you can see which cost is pending, what evidence supports its allocation, and who needs to confirm it. If reporting work is already taking a substantial portion of an AM’s week, tightening the underlying data can reduce the time spent reconciling numbers before they can be discussed with a client.

The Reporting Agent can pull performance data and draft the monthly report and account manager email summary for review. It works alongside expense allocation, not instead of it. Cleaner project costs give reports a better view of delivery economics. The Account Health Agent can then help flag client risk and opportunity, so the team can act on patterns rather than spend its time assembling them.

Where to start without adding another spreadsheet

Start with one expense category or one group of projects. Choose a workflow where the agency has enough transactions to test, but where a wrong allocation can be reviewed before it affects client billing or a closed month. Common starting points include stock assets, travel, freelance production costs, and recurring software charges.

Then agree on four things:

Which project and internal codes are valid.
What evidence is enough for an automatic recommendation.
Which expenses always require approval.
Who owns each exception type and how quickly it should be resolved.

Keep the first version narrow. If receipt matching is unreliable, fix that before adding more allocation rules. If projects lack consistent codes, clean up the project list before asking an agent to choose among near-duplicate names. If approvals stall because no one owns a shared expense, assign a role rather than sending more reminders.

The point is not to automate judgment away. It’s to reserve judgment for the cases where the evidence is incomplete or the cost matters enough to review, while routine matches move through with a clear record.

What an Omni Audit gives you

An Omni Audit can help your team examine where operational work gets stuck and what to address first. For an agency expense workflow, that means tracing a transaction from card charge and receipt through project coding, approval, and reporting. We look at the systems involved, the handoffs between finance and account teams, and the exceptions that keep coming back.

The session can help your team consider the current workflow, automation opportunities, and practical next steps. No deck. The aim is to identify which parts are ready to automate, which need better rules or data, and where a person should stay in the approval loop.

See Omni for marketing and creative agencies for the vertical-specific audit. You can also book your 60-min Omni Audit and bring a recent expense exception or month-end reconciliation example.

For an agency with possible annual leakage across operational issues, the useful question isn’t whether every dollar can be recovered through expense allocation. It can’t. The useful question is how much time and margin are being lost because costs aren’t attached to the right work, and what a cleaner process would let your team see sooner. Start with one real month of transactions, trace the exceptions, and build from what the evidence shows.

See the AI audit for marketing and creative agencies to take that first step.

Reconcile Agency Expenses to Client Projects