Working capital

An income statement records a sale when it is earned. A bank account records it when the money lands. Between those two events the sale has to exist somewhere, and it exists as a balance on the balance sheet. That is all working capital is: the accumulated difference between when things are recorded and when they are paid for.

Three balances carry it. Receivables hold sales made and not collected. Inventory holds goods paid for and not sold. Payables hold purchases received and not paid, which is the same mechanism running in the business's favour. Add the first two, subtract the third, and the total is money that is in the business and not available.

Each balance is one multiplication, driven by the month's own activity. Receivables scale with revenue because that is what customers owe against. Inventory and payables scale with cost of sales because that is what was bought.

The last line is the one that reaches the cash flow statement. It is the change in the three balances, not the balances themselves, because a business with a large but stable receivables book is not consuming cash. It is only growth in the balances that takes money out.

No credentials of any kind are involved. If an example ever needs one, the string is the word placeholder.

Days sales outstanding is how long the average pound of revenue sits as an invoice before it becomes money. Marchmont's is 75, so the balance is revenue x 75 / 30, two and a half months of sales held as paper. In month 1 that is £1,000,000, which is already larger than the opening cash balance.

What it decides. Credit terms, how hard collections are chased, and which customers get a credit limit at all. Where judgement enters. DSO is an average and averages hide the shape. Ninety-nine laboratories paying on the nail and one large teaching hospital nine months late produce the same DSO as everybody paying moderately late, and those are entirely different problems with entirely different fixes. A single figure should always be followed by an ageing profile.

Days inventory on hand is how long stock sits before it becomes a sale. Marchmont's is 60, computed on cost of sales rather than revenue because stock is carried at what it cost, not what it will fetch. In month 1 that balance is £464,000.

What it decides. Whether a laboratory that orders on Tuesday gets an instrument this month or next. Two months of finished stock is a promise of availability, and availability is part of what a B2B buyer is paying for. Where judgement enters. The right level is a bet about demand nobody can settle in advance. Hold too much and the money is idle and the instruments may be superseded. Hold too little and the order goes to a competitor who had one on the shelf. Neither error appears in the accounts as an error.

Days payable is how long the business holds its own suppliers' money before handing it over. Marchmont's is 30, against a balance of £232,000 in month 1. Every pound sitting in payables is a pound of the business's operations that a supplier is currently financing at no charge.

What it decides. How much of the cycle somebody else pays for. At 30 days against a 135 day journey, Marchmont's suppliers cover less than a quarter of it. Where judgement enters. Goodwill has no ledger account. Terms are extended by negotiation or by simply paying later and waiting to see what happens, and the second route works until the month a supplier is deciding who gets the last shipment of a scarce component.

The cycle is DSO plus DIO less DPO. For Marchmont that is 75 plus 60 less 30, which is 105 days. It is the number of days between the business's money leaving and the same trade's money returning, and it is the single most compact description of how much cash a business needs simply to operate at its current size.

The arithmetic runs in both directions. A retailer who sells for cash, holds stock briefly and pays suppliers on long terms has a negative cycle: the customer's money arrives before the supplier's invoice falls due, so growth generates cash instead of consuming it. A subscription business billed annually in advance is in the same position. Marchmont is the opposite case, and the sign of the cycle is most of what determines whether growth needs funding.

The practical use is that the cycle multiplied by daily sales gives a rough size for the working capital a business must carry. Cut the cycle and that requirement falls, in proportion, without a single sale being added.

Only the change in the three balances touches cash. A business selling the same amount every month has a large receivables book that does not move, and a balance that does not move takes nothing out of the bank: the collections coming in match the invoices going out. Marchmont grows 4% a month, so each month's balances are larger than the last, and the difference is funded out of the bank balance.

That is the mechanism behind the whole course. Growth does not consume cash because growing is expensive. It consumes cash because a larger business needs a larger amount of money permanently parked in invoices and stock, and the increase has to be found before the extra profit arrives.

Operating cash flow is net income plus depreciation less the working capital movement. Marchmont's is negative in every month from month 2, and it climbs steadily back towards zero without ever reaching it: minus £23,481 in month 2, minus £5,796 in month 10, minus £427 in month 12.

It climbs because the two terms grow at different speeds. Net income rises from £17,732 in month 2 to £63,370 in month 12. The working capital movement rises from £49,280 to £72,946 over the same stretch. Profit is catching up. On this trajectory it eventually overtakes, which is session 9 and is the reason a twelve month window is not long enough to judge the business.

Two fields on every row, and confusing them is the commonest mistake in this subject. workingCapital is a level: receivables plus inventory less payables, at a moment. deltaWC is a flow: how much that level moved this month, and it is the only one of the two that appears in a cash flow statement.

A business can carry an enormous working capital level and generate perfectly healthy cash, provided the level is stable. It can also carry a small level and haemorrhage cash, if the level is doubling. The question is never how much working capital there is. It is which direction it is moving and how fast.

None of the three dials touches a line on the income statement. Collecting sooner changes when a recorded sale becomes money, not whether it was earned. Paying a supplier later changes when a recorded cost is settled, not what it cost. Holding less stock changes how much is sitting in the warehouse, not what was sold or at what margin. So revenue, cost of sales, gross profit, operating expenses, EBITDA, EBIT, interest, tax and net income are all identical across the base case and all three levers.

That is the demonstration to keep. Cash at month 12 moves from minus £47,261 to £168,521 on the first lever, a swing of more than two hundred thousand pounds, and the twelve month profit figure is £445,631 in both runs. Anyone who believes profit and cash are broadly the same thing can be talked out of it in one console command.

Every lever on that diagram is a change somebody else has to accept, and the model has no opinion about whether they will.

• Collecting in 45 days rather than 75. Those terms are in signed contracts. Changing them means going back to laboratories that agreed 75 days and asking for something worse, in a market where a competitor is still offering 75. Some will agree. Some will use it as an opening to renegotiate price. Some will buy elsewhere next time.
• Paying suppliers in 60 rather than 30. This is spending goodwill that has no balance. It works until the quarter a component is scarce and a supplier is choosing which customer gets the allocation, and that cost arrives long after the cash benefit was banked and is never attributed to it.
• Holding 30 days of stock rather than 60. Half the buffer means half the protection against a demand spike or a late delivery. A stockout costs an order, and possibly the customer. The order that was never placed does not appear in any row of any model, which is exactly what makes this lever look free.

The model prices the cash precisely and the costs not at all. A recommendation built on the cash figure alone is a recommendation with one side of the ledger missing, and saying so out loud is part of presenting the number.

One position. An unpaid invoice is an interest-free loan the supplier never agreed to make. A business that does not manage its terms is financing its customers' balance sheets out of its own, and the cost is real: it is the interest on the debt drawn to cover the gap, or the growth foregone because the cash was not there. Terms should be set by whoever owns the balance sheet and enforced, and a sales team that hands out 90 days to close a quarter has spent money it does not control.

The other position. Terms are part of the price. A laboratory comparing two instrument suppliers is comparing total cost including when it has to pay, and 75 days is part of what Marchmont is selling. Pull it to 45 and some customers pay the same, some demand a discount that gives back the benefit, and some quietly go elsewhere. The ones who go elsewhere never appear as a working capital line. They appear as revenue that was never booked, which no model can distinguish from a market that simply did not buy.

Both sides are arguing about a quantity nobody in the room has measured, which is how much business the terms are buying. The useful discipline is to say which position you hold, say what evidence would change your mind, and notice that the £168,521 on the diagram is the upper bound of the benefit rather than the benefit.