Asked how much cash the business generated last year, three people at the same table will frequently produce three different figures: the finance lead subtracting capital expenditure from EBITDA, the accounting side adding depreciation back to net income, the owner naming the difference between the opening and closing bank balance. Each answer is internally coherent, none of the three is deliberately misleading, and none of the three measures the same quantity. The spread between them does not narrow as the company grows. It widens, as legal entities multiply, as the investment programme moves from a single annual decision into phased commitments, and as supplier payment terms diversify across categories. That widening is a function of definitional vacancy rather than of accounting quality, which is precisely why it survives an otherwise clean audit.
At the review desk the opening question is rarely about magnitude. What is asked instead is which definition produced the number, whether that same definition was applied without variation across all three prior years, and from which line of the audited statements the calculation departs before arriving, through a sequence of adjustments, at the figure now on the page. The answer, more often than not, reveals that the definition is less an established internal construct than a presentation item assembled for the meeting itself. Whether free cash flow exists as a verbal assertion or as an institutional computational discipline becomes evident within the first half hour of diligence, and it sets the register of everything negotiated afterwards.
The mechanism beneath the gap is not inattention but the economics of measurement. A company measures what it manages day to day, and what is managed is collection, the payment calendar and the bank balance. The bank balance is concrete, observable every morning and free to verify; functioning as a mental anchor, it becomes the reference point around which all intuition about cash generation is organised. Free cash flow, by contrast, is not directly observable — an abstract quantity assembled from several components, none of which can be confirmed by looking at a single screen — and, being unobservable, it is perpetually deferred to the next stage of institutionalisation. Because profit and cash are held in separate mental accounts, movement in working capital, which appears nowhere on the income statement, drops out of the cash conversation altogether.
Under a particular configuration this shortcut is genuinely functional and entirely rational. Where collection and payment pass through a single approval line, where investment decisions are taken a handful of times a year by the same person, and where inventory turnover moves seasonally within a narrow band, the bank balance is in fact an adequate indicator; the cost of maintaining a separate cash bridge exceeds the information it yields. The difficulty lies not in the shortcut but in its persistence after the conditions that justified it have changed. Once entities multiply, once capital expenditure is phased, once supplier terms differentiate, and once growth converts working capital into a cash-consuming line, the bank balance no longer reflects the company's capacity to generate cash — it reflects the sequencing of payments in the final quarter.
On the documentation dimension the typical position is this: the definition lives in a spreadsheet on one person's machine, the formula has shifted quietly from year to year, no version history exists, and the reconciliation step tying the model to audited statements has never been written down. Placed in a data room, a file of that kind is classified not as verifiable evidence but as an assertion requiring verification. The distinction is not merely procedural, since an undocumented calculation lengthens the diligence timetable and, in lengthening it, consumes the seller's negotiating time. Institutional memory shows its weakness here in its most tangible form: with the person who performs the calculation on leave or departed, the figures for the preceding three years cannot be reproduced.
The first channel through which the institutional cost travels is the counterparty's reconstruction of the definition. A reviewing party building a number for its own investment committee rebuilds it from source documents, and does so on conservative assumptions, with the consequence that any item of uncertain classification is predictably assigned to the cash-consuming side. The most visible instance is the treatment of capital expenditure: where the boundary between maintenance and growth investment is not documented at the asset level, the whole of capex is taken as maintenance within normalised cash flow. Presented as a single line adjustment, the effect is in fact multiplicative, because it lowers the base to which the multiple is applied, and it commonly exceeds a full year of operational improvement.
The second channel is that the gap tends to surface not in the price but in the closing mechanics. The normalised working capital peg is conventionally established from twelve months of monthly data; absent monthly data, or with only year-end snapshots available, the reference level is calibrated close to the highest observed point and the completion adjustment moves systematically against the seller. The same ambiguity pushes earn-out structures away from EBITDA and towards cash conversion, raises the escrow ratio, and broadens the representation and warranty package under the working capital heading. On the lending side the transmission is more direct still: where it remains unclear which cash definition underpins the DSCR calculation, the covenant package is typically calibrated within a tighter band.
The third channel is ownership and continuity. In most companies no named role carries responsibility for the cash conversion cycle in its entirety — the commercial side manages collection terms while being measured on revenue, procurement manages payment terms while being measured on unit price, operations sets inventory levels while being measured on delivery performance. All three behave rationally against their own objectives, and the total length of the cycle appears on no one's scorecard. The continuity question sharpens the point further: if the accuracy of the cash forecast rests on one individual's recollection of seasonality and customer payment behaviour, that capability belongs to the individual rather than to the company, and it is priced in diligence as founder dependency rather than as financial performance.
What neutralises this tendency is architecture rather than individual discipline, and it separates into four components. The first is a versioned and approved definition memorandum that begins at a specified line of the audited statements and records each adjustment step together with its rationale. The second is a monthly cash bridge reconciled against bank movements, functional to the extent that it allocates the difference between profit and cash into named items every month rather than annually. The third is a capital expenditure register in which each commitment is classified as maintenance or growth; the decisive detail is that classification occurs at the point of proposal rather than at the point of approval, since a classification produced afterwards reads as defence rather than as reasoning. The fourth is a review rhythm in which the variance between the thirteen-week cash forecast and actual outturn is measured, with the variance itself treated as a performance indicator.
BEIREK begins by fixing the definitional layer: the cash flow definition, the adjustment items and the reconciliation steps are committed to writing, carry an approval date, and may be altered in later periods only against a recorded rationale. The monthly cash bridge and the capital expenditure classification register are then put into operation, and ownership of the full conversion cycle is attached to a single named role, with collection, inventory and payment terms written jointly into that role's measurement set. The detail that proves decisive in practice is that this rhythm should have been running for at least two or three quarters before the data room opens; a bridge assembled retrospectively during diligence is typically read as a reconstructed presentation rather than as an established system, and it does not relieve the verification burden.
The information free cash flow actually carries as a diligence item is not how much cash the company generated last year, but whether the company can describe and reproduce its own cash generation independently of the people who manage it. Where that capacity exists, the number itself moves onto negotiable ground, because the discussion proceeds over assumptions rather than over definitions. Where it does not, the company comes to the table not simply with a computational gap but with the counterparty's conservative definition already installed, and the cost of that definition is usually visible less in the price than in the structure through which the price is paid.