Profitable on Paper, Tight in the Bank: Cash Flow vs Margin for Agencies

I have watched a finance lead present a confident, profitable month to a leadership team on a Monday and quietly tell me on the Wednesday that we might need to delay a supplier payment to clear payroll. Both statements were true. The month was profitable. The bank was tight. Nobody had done anything wrong.

This is the most common financial surprise at a growing agency, and it catches good operators because the two numbers that describe it look like they should agree. Profit and cash feel like the same thing — money the business made. They are not the same thing, and the distance between them is where a lot of agencies get into trouble in exactly the years they are growing fastest.

This piece is for the CFO, CEO or owner who can read a P&L, sees a profitable line at the bottom, and still feels the monthly tightness in the bank account that nobody quite explains. The short version of the profit vs cash flow problem is that margin is an accrual concept and cash is a timing concept, and the gap between them is mostly a billing-and-collection problem you can measure and shrink.

Margin is an accrual idea; cash is a timing idea

Margin answers the question did this work make money. It records revenue when the work is earned, not when the invoice clears, and it records cost when the cost is incurred, not when the supplier is paid. That is the accrual principle, and it is the right way to judge whether a project or a month was profitable. (If you want the full definition of what real project margin includes — labour at real cost, allocated overhead, unbilled scope — that is in what project margin actually is.)

Cash answers a different question: do we have the money right now. It records money when it moves. A client paying a £60k invoice is a cash event on the day the funds land, regardless of when the work was done. Paying your team is a cash event on payday, regardless of which projects those people were on or whether those projects have been paid for.

Both are honest. They simply run on different clocks. Margin runs on the clock of when value changed hands. Cash runs on the clock of when money changed hands. At a manufacturer those clocks are reasonably close together. At a services business that delivers continuously and bills in arrears, they can be a month or two apart, and the gap widens precisely when the business is growing — because growth means more work delivered today that will only turn into cash later.

The mistake I made for years was treating the P&L as the whole picture and the bank balance as a thing that would sort itself out if the P&L was healthy. It does not sort itself out. A profitable agency that bills slowly and grows quickly can run out of cash with a perfectly good margin. The margin is real. So is the overdraft.

Where the gap opens

The distance between profit and cash is not random. It comes from a small number of recurring, measurable timing items. If you know the five, you can forecast the gap instead of being surprised by it.

Work in progress: delivered, not yet invoiced

The biggest one at most agencies is work in progress. Your team logs hours against a project this week. The value of that work is real — it counts towards margin the moment it is delivered. But you may not invoice it for weeks, because the contract bills at a milestone, or monthly in arrears, or at the end of a phase. In the meantime you have paid the salaries that produced it.

WIP is the pile of delivered-but-not-yet-invoiced work sitting between your team's effort and your bank account. The faster you grow and the longer your billing cycle, the bigger that pile gets. It is an asset on paper and a hole in the bank.

Milestone and stage billing lag

Fixed-price builds that bill at stages — 30% on kick-off, 40% at a milestone, 30% on delivery — concentrate cash into a few dates while the cost of delivery spreads evenly across the whole project. Between the 30% deposit and the milestone payment, you might fund three months of payroll on a single project with no cash coming in against it. The project is profitable. The middle of it is cash-negative. (The structural margin and cash trade-offs between fixed-price and time-and-materials are worth their own read: fixed-price vs time and materials margin.)

Deferred revenue on annual retainers

Annual retainers paid up front are the one item that runs the other way, and they hide the problem. A client pays twelve months of support in January. That is a large, welcome cash inflow. But only one-twelfth of it is earned revenue in January — the rest is deferred revenue, a liability you owe the client in service across the year. If you read the cash and think it is profit, you will spend money you have already promised to work off. Deferred revenue flatters cash today and constrains it tomorrow.

Payroll goes out before client payment comes in

Payroll is the least forgiving line in an agency. It leaves on the same day every month, in full, whether or not a single client has paid. People are roughly 70–80% of an agency's cost base, and that cost is paid in near-real-time while the revenue it produces is collected in arrears. The structural position of a services business is: pay for the work now, collect for the work later. That is working capital, and growth consumes it.

The VAT and tax timing trap

The quietest one. VAT is collected on your invoices and held until the quarterly return is due. For a quarter or so it sits in your account and feels like cash you have. It is not your money; it is the tax authority's money you are holding. The same is true of corporation tax accruing against your profit and of any PAYE timing. When the VAT payment lands — often £80k–£150k at once for a mid-sized agency — it can turn a comfortable month into a tight one overnight if you have been mentally counting that balance as your own.

A profitable month that goes cash-negative

Here is the worked example I use to make the point. The numbers are anonymised and scaled to round figures, modelled on a ~50-person agency.

Start with the P&L for a single month:

• Revenue recognised (work earned in the month): ~£600k
• Cost of delivery (labour at real cost): ~£380k
• Overhead (rent, non-billable staff, software, leadership): ~£120k
• Operating profit on the P&L: ~£100k

A genuinely profitable month. ~17% operating margin. Leadership is pleased on Monday.

Now the cash for the same calendar month, which is a completely different statement:

• Cash collected from invoices issued ~45 days ago: ~£430k. (The month's £600k of earned revenue is not collected this month; some of it is not even invoiced yet — it is WIP.)
• Invoices actually issued this month: ~£400k. The other ~£200k of earned revenue is sitting in milestones and monthly-in-arrears cycles that have not billed.
• Payroll paid: ~£350k, in full, on payday.
• Other operating costs paid: ~£110k.
• Quarterly VAT payment due this month: ~£95k.

Net cash movement: £430k collected − £350k payroll − £110k opex − £95k VAT = −£125k.

So the month is +£100k on the P&L and −£125k in the bank. Nothing is wrong. No client has defaulted, no project is loss-making, the margin is healthy. The agency has simply delivered £600k of value, collected £430k of older invoices, paid its people in full and handed the VAT authority a quarter's worth of held tax in the same thirty days. A swing of ~£225k between the two numbers, in a profitable month.

Run three months like this back-to-back during a growth spurt — more work delivered, more WIP, bigger payroll, the VAT quarter landing in the middle — and a profitable agency is having an uncomfortable conversation with its bank. Not because the business model is broken. Because the timing is.

This is what a continuous margin view does and does not help with. It will not pay your VAT bill. What it does is shrink the part of the gap you actually control: it shows you the WIP pile in real time, so you can see how much delivered-but-uninvoiced work is sitting between effort and cash before it becomes a surprise.

Watching both numbers without drowning in spreadsheets

The instinct, once you have seen a gap like this, is to build a heavy thirteen-week cash model and update it forever. You do need a cash forecast. But most of the practical improvement comes from a small number of disciplines that shrink the gap rather than just describe it.

Watch margin and cash as two separate, equal numbers. Neither one is the truth on its own. Margin without cash is a business that looks healthy until the day it cannot pay people. Cash without margin is a business that feels fine because a retainer landed up front, while the underlying work loses money. Leadership reviews should show both, side by side, every week.

Shrink the WIP-to-invoice lag with billing cadence, not heroics. The single biggest lever most agencies have is how fast delivered work turns into an issued invoice. Monthly-in-arrears with a hard close on the last working day beats "we'll bill when the phase wraps". Bill milestones the day they are hit, not the week after someone remembers. Every week of lag you remove is a week of payroll you no longer have to pre-fund.

Tighten collection, not just invoicing. Issuing the invoice is half the job. Payment terms, deposits on new work, and a named person who chases overdue invoices on a schedule are what actually move cash. A 45-day collection cycle and a 30-day one are a month of working capital apart on the same revenue.

Treat deferred revenue and held tax as money you do not have. Annual-retainer cash and the VAT balance are the two amounts most likely to be mentally double-counted. Ring-fence them. The number you can spend is the number after you have set aside what you owe in future service and in tax.

Make WIP visible continuously. This is the part where reading your Jira worklogs in real time helps directly. The same data that tells you a project's live margin also tells you how much delivered work is sitting un-invoiced right now. A few of the weekly Jira queries that surface stalled and drifting work — the ones in four Jira queries every agency CFO should run — are also early-warning signals for cash, because a ticket that is accumulating cost without moving towards a billable milestone is WIP that is quietly growing.

The goal is not to make cash and margin equal. They measure different things and they should never be expected to agree in a given month. The goal is to know the size of the gap in advance, control the part of it that is yours to control — the billing and collection cadence — and stop being surprised by the part that is structural.

A profitable agency that runs out of cash did not have an accounting problem. It had a timing problem it could not see. If you would like to see your delivered-but-uninvoiced work and your live project margin on your real Jira data — the two numbers that sit on either side of the cash gap — the 15-minute Saldo demo runs on your own instance, read-only, and we do not ask for card details up front. Margin tells you the model works. Watch cash too, so the model gets the chance to.