Finance: Lost Revenue, or Just Late Revenue?

Thursday morning of the second week. Claire Beaumont has four files open and a one-line request from Ravi: what does each option cost, and when? The when is doing most of the work in that sentence, and she is the only person in the building who will treat it as the important half.

Everyone else has been describing this crisis in ovens. Claire’s job is to convert units into money, and the moment she starts she meets the question this module turns on. Is Calder about to lose $5.3m of revenue, or to receive $5.3m rather later than planned? Those are wildly different sentences, and the honest answer is mostly the second one, with an expensive exception.

What lands on Claire’s desk

From Grace come two lines that appear in no operational plan: $1.85m of work in progress — 240 ovens built to one component short, parked on the floor — and about $115,000 of overtime to lift the recovery rate once boards flow. With them come dates: build-ahead 20 April to 22 May, full rate resuming 26 May.

From Yusuf comes demand restated from 960 units to 826, range 790 to 870, the reduction sitting almost entirely outside the protected list — better described, not smaller. From Hannah come the price of each lever and the condition attached to it. From Tomasz comes the phased allocation, which bakes quarter-end into a schedule rather than a forecast.

Two things are fixed and not hers to reopen: the 520 boards are the 520 boards, and Ravi’s objective from module two — protect first the relationships that cannot be bought back at any price, recover revenue second, spend only where it buys something lasting beyond the quarter. Claire is not being asked whether to spend, but what each choice costs, in which currency, and in which reporting period.

What a finance business partner actually does

A finance business partner sits inside the operational meeting rather than receiving its minutes: traditional finance reports last month, partnering prices what is about to happen while it can still be changed. The trap is one of two caricatures — the accountant who blocks spending because it lands in the wrong quarter, or the enthusiast who approves anything with a good story.

The software on Claire’s desk

The ERP supplies the figures nobody can dispute, which is the foundation of finance’s authority in the room. The scenario model in Excel is where the crisis is actually thought about: change the assumption about how much deferred revenue never returns, and watch the answer move — that sensitivity is more useful to a board than any single number.

The planning tool carries the cash view, and this is the screen that separates finance’s anxiety from operations’ calm: profit and cash behave differently when $1.85m of nearly-finished ovens sit on a floor. The board paper does the job that gives this desk its purpose — putting an expedite fee, a tooling investment, a redesign and a penalty clause onto the same basis, so the decision is made on merit rather than on which option happened to be quantified most precisely.

The decisions

Lost revenue, or deferred revenue

Most of this money is not destroyed. It moves. A hotel group that waits three weeks still buys the ovens; the invoice is raised in June instead of April. Claire says so out loud in the first meeting, because half the room is quietly assuming $5.3m has evaporated, and panic about destroyed revenue costs more than the disruption itself.

That is genuinely good news for the year. Calder turns over $50m, and ovens delivered in June rather than April barely move an annual outcome. It is still a serious problem, because reporting periods are real. The quarter closes in week nine, which on Grace’s dates falls in the middle of the build-ahead, when the factory is busy and invoicing almost nothing. A large slip across that boundary changes what the company reports, which drives bank covenants, board confidence and the tone of the next three months. “It all comes back next quarter” is true, and does not survive contact with a covenant test.

Then the part people miss. A fraction of deferred revenue quietly never comes back, because some customers cancel rather than wait. Bellwether’s dry dock cannot move, so lateness there is not a delay but a cancellation. Nordfoods is openly dual-sourcing. Acme Bakery can buy a competitor’s oven on Thursday. The comfortable assumption — everything returns — is wrong in one direction only.

So Claire refuses a single blended percentage across the book, because the leakage is concentrated in named accounts rather than spread evenly. She estimates it customer by customer, with a cancellation probability against each, printed on the face of the paper where it can be argued with.

Cash, not profit

Here is the sentence that separates finance from everyone else: a profitable company can be in serious trouble during a disruption, and the trouble arrives suddenly.

Grace’s build-ahead is the clearest example. Those 240 nearly-finished ovens are $1.85m of materials and labour Calder has already paid for, with nothing invoiced, because nothing has been delivered. Meanwhile the overtime is paid in the month it is worked, the air freight when it is booked, the broker’s boards on the spot, Kestrel’s tooling on order — all cash out, now. The cash in arrives sixty days after a delivery that has itself slipped by weeks, so an invoice due in June is collected in September.

Working capital is simply the money tied up in the gap between paying for things and being paid for them. In normal trading it sits at a steady level and nobody thinks about it; a disruption stretches the gap at both ends at once, and the peak lands almost exactly at quarter end.

This is why finance directors get nervous at the moment operations relaxes. To Grace, 240 banked ovens are recovery capacity — the right call, and Claire does not argue with it. To Claire they are $1.85m converted from cash into floor space at the point in the year with the least incoming cash. Neither view is wrong; what is wrong is only one of them being in the room. So her decision is not to block the build-ahead but to model the cash trough — how deep, which week, whether the facility covers it — and tell Eleanor before it happens rather than after. Profit is an opinion formed over a quarter; cash is a fact that arrives on a Tuesday.

Pricing each lever as an investment, not a cost

Operations brings Claire four price tags. She refuses to treat any of them as a cost, because a cost is something you minimise and an investment is something you evaluate against what it returns.

Expedite is the easy one, and instructive because it is easy. $85,000 buys about two weeks: roughly 160 ovens, $1.9m of revenue, $630,000 of gross profit. Twenty-two dollars back for every dollar spent is not a close call — and the scale of that margin is the point, because it makes the price irrelevant. At these odds the only thing worth arguing about is whether Vantor’s date holds, since air freight booked against a date that slips still saves two weeks, just two weeks too late for the customers it was meant to save. Claire’s contribution is not approval but reframing: a confidence decision, so the probability goes in the paper and a trigger goes on the money.

Kestrel is the hard one. $180,000 of tooling and $140 a board more, for a second source that cannot arrive for ten weeks — bought, in other words, for a crisis it cannot reach. On this quarter’s return it scores zero, and a finance function that judges it that way will kill it. It is insurance, and insurance is evaluated not against this year’s return but against the cost of the event multiplied by how often the event happens. This shock puts roughly $1.7m of gross profit at risk; if a sole-source failure of that size comes round once every five years, the expected annual cost is near $340,000 — against a premium of perhaps $60,000 to $90,000 a year, being the tooling plus the price gap on the modest volume needed to keep a second supplier live. Nobody asks whether their building insurance paid back this year.

The redesign at about $250,000 follows the same logic, with one advantage: a premium paid once rather than annually, because it removes the single point of failure instead of hedging it.

Ravi’s third clause now does real work: Kestrel and the redesign buy something lasting beyond the quarter, expedite does not and is justified on straight return instead. Finance’s job is to make the board compare like with like — mitigation judged on payback inside the window, capability on expected loss avoided over years. Put them in one column headed “cost” and the insurance always loses, because it is the only line with no benefit this quarter.

The penalty arithmetic, and what it does not capture

Harlow Hotels’ contract carries liquidated damages at 0.5% a week, capped at 5%, on an order worth $2.52m. That yields two numbers of unusual quality: each week of lateness costs exactly $12,600, and the worst case is exactly $126,000, reached after ten weeks.

This is the only cost in the entire crisis that is precisely knowable, and its precision is dangerous. Put $12,600 a week on a slide beside a row of estimates and it wins every argument it enters — not because it is the largest number, but because it is the only one nobody can dispute. The distortion is systematic and it runs one way: customers with penalty clauses look expensive to disappoint, and customers without them look free.

Nordfoods has no penalty clause. It is also worth roughly $1.4m of revenue a quarter and, at the Compact’s 26% margin, some $370,000 of gross profit a quarter, and it is openly dual-sourcing — so lateness there costs not a penalty but a share of that volume permanently. Even a fifth of it moving to their second supplier is worth more, every year, than Harlow’s entire penalty cap once. Bellwether carries no damages either; it carries a cliff, because missing the dry dock cancels roughly $720,000 of order outright with no repeat for five years. St Chad’s costs nothing in cash and risks a contract place at renewal.

So Claire does the thing finance is often reluctant to do: she puts a number, however rough, on every un-penalised loss, and labels it an estimate. A wide range with its assumptions named beats a blank, because a blank is read as zero. Left unpriced, the precise number wins by default — and Calder pays $85,000 of air freight to protect $12,600 a week while quietly losing an account worth ten times that.

Where this goes wrong

Two weeks into a disruption, most companies are running several versions of the same arithmetic: operations has a spreadsheet of units, sales one of revenue at risk in their accounts, finance a third built on different assumptions about what returns. All three are defensible, none agree, and the board meeting is spent reconciling them instead of deciding anything — the most expensive way for a company in a shortage to spend a week.

The subtler version is worse: a confident figure that quietly assumes every deferred order comes back, because nobody wants to be the person forecasting cancellations. It holds until the cancellations arrive, taking the credibility of the whole finance view with it. The defence is unglamorous — one model, owned by one person, assumptions visible on the page, and other desks welcome to argue with them rather than keep rival copies.

What Claire hands on

To Ravi and the board goes an agreed cost-of-options paper: each lever with its price, its return, the period that return lands in, and the condition it depends on — expedite as a confidence decision with a trigger on the money, Kestrel and the redesign priced as insurance, the broker batch shown as $44,000 at risk until it passes test. Crucially it is one paper, and every desk has already seen its own numbers inside it.

To Eleanor goes the cash forecast: the depth and timing of the working capital trough, driven by $1.85m of WIP, immediate outflows and receipts landing sixty days after deliveries that have already moved. It turns “we will be fine over the year” into a specific week that needs facility headroom.

To Marcus, for module nine, goes the number nobody has yet produced: the true cost of each week of delay, customer by customer. Harlow’s clean slope to a ceiling; Bellwether’s zero, zero, zero, then a cliff; Nordfoods’ near-zero fortnight followed by a permanent step down in annual volume; St Chad’s contract place; Peninsula’s discount; Acme’s $6,000 and a story. The shapes are all different, and that is the point: he cannot rank customers by order size, because the cost of disappointing one has almost nothing to do with how much it buys.

None of that is advice, and none of it is permission. What finance owes the other desks in a crisis is a single agreed set of numbers, so that sales, operations and the board spend the next eight weeks arguing about the decision rather than about whose spreadsheet is right.

Spot the decision

Read each situation and decide how a finance business partner should handle it, then tap a card to check.