In a diligence process the first request concerning gross margin arrives in almost identical form every time: three years of margin, broken out by product or project line. The company side usually meets that request without much friction — a schedule is assembled, the lines are opened row by row, the ratios are entered. Divergence begins with the second request, which asks that the schedule be reconciled to the audited financials, and once that reconciliation is attempted the aggregate cost carried in the schedule and the cost of goods sold reported in the income statement frequently fail to meet. The gap is not an error on either side; it is the consequence of two boundaries having been drawn by two different hands for two different purposes, neither of which was ever obliged to agree with the other.
The same pattern is observable inside the company long before any external party arrives. A line margin quoted from memory during a pricing discussion may differ materially from the figure that emerges from the accounting records, given that freight, installation, warranty rework, packaging, and the travel cost of field personnel are frequently treated as cost on the pricing side while sitting in operating expense on the books. A business can run on both numbers for years, each internally consistent within its own use, without any contradiction surfacing, so long as the two populations of users never compare notes. The difficulty begins when a third party sees both at once and asks which of the two describes the business being sold.
The mechanism sits precisely here: margin is a defined boundary before it is a measured quantity. Direct labour, the cost of idle capacity, depreciation on production equipment, the quality control headcount, post-sale correction work, internal logistics, and — in project-based businesses — site overheads can each be brought inside cost of goods sold or left outside it on defensible grounds. Absent a written decision recording where the line runs, the resulting ratio is less a performance indicator than a derivative of how the chart of accounts happened to be constructed years earlier, often by a bookkeeper responding to a tax question rather than a commercial one. The existence question in a review is therefore not whether the ratio is calculated, but whether its definition formally exists.
The absence of a written definition is rarely negligence; it is a shortcut that was rational at a particular scale. Where price is set by the founder or by a single commercial director, formalising margin knowledge adds little, because the decision-maker already carries what each type of work leaves behind and which combination of term and discount remains acceptable for which customer. The shortcut is cheap and fast, and it holds while the number of decision-makers remains one. The problem lies not in the shortcut itself but in its persistence as the product count, the customer count, and the number of people authorised to quote all increase; beyond that point margin ceases to be an input at the moment of decision and becomes a result learned when the year closes.
What is typically observed on the measurement dimension follows directly from that retrospective computation. Where margin is known only annually and in aggregate, the average conceals the distribution beneath it: two portfolios producing an identical average may look entirely different, one clustered within a narrow band, the other composed of high-margin work offsetting work priced below cost. In the second configuration growth pulls margin down systematically, since incremental volume generally arrives from the segment where price pressure is highest, and because the cross-subsidy is not visible in any report that segment is fed on the assumption that it is profitable. Without monthly margin tracking broken out by customer and product, the erosion becomes apparent only when total profit falls, by which time the contracts have already been signed.
The institutional cost crystallises during the financial validation phase of the review. A quality of earnings exercise conducted on the buy-side treats any classification unsupported by documentation in the conservative direction; items open to question migrate from operating expense into cost of goods sold, not out of hostility but because the reviewing party’s obligation is to present its own investment committee with a defensible floor rather than to confirm the seller’s presentation. That prudence runs in one direction only, and it pushes the normalised margin below the level the company put forward. The reduced margin then affects more than the historical record, since it also resets the starting point from which every projection over the plan period is built.
From there the effect reaches valuation through several channels at once. Where the margin assumption embedded in the projection cannot be supported by the historical series, plan-period profitability is discounted and the multiple is applied to the lower base, so that the two adjustments compound rather than merely add. The second channel is transaction structure: an unverifiable margin normalises the deferral of part of the consideration into an earn-out, and the earn-out trigger is very often set on precisely the gross profit metric whose definition remains contested, which manufactures a fresh dispute surface over calculation methodology after closing. The third channel runs through the scope of representations and warranties and the escrow ratio, since narrower coverage of the financial statements is typically compensated by a higher retained amount.
The ownership and continuity dimensions converge on the quietest and most decisive question in the entire exercise: who is able to refuse a price below a given threshold, and on what authority that refusal rests. In structures where pricing approval collapses into a single signature, where the threshold is unwritten, and where exceptions leave no record, margin discipline is not a system but one person’s stamina. To test this, a counterparty will commonly select the lowest-margin jobs of the trailing year and ask how those prices came to be set; where the answer rests on recollection rather than on a chain of calculation, the continuity dimension is treated as unmet and the valuation carries a founder-dependency discount that no amount of historical profitability offsets.
The intervention that corrects this picture operates at the level of architecture rather than awareness, and it separates into four components. The first is binding the cost of goods sold boundary to a written classification decision, so that which item sits inside, which sits outside, and on what reasoning are fixed in a short document that also commits the company to consistency across periods. The second is moving the moment of computation from year-end to the point of quotation, so that every offer is produced alongside a record carrying its own expected margin. The third is bridging the variance between quoted and realised margin on a monthly rhythm, broken out by customer, product, or project. The fourth is assigning the authority to approve below-threshold pricing to a named role and recording every exception with its rationale.
BEIREK’s work in this area typically begins with the classification decision itself, since every subsequent measurement layer is constructed on that boundary: cost items are taken one at a time, the boundary decision is written together with its reasoning, and prior periods are recalculated on the same basis so that a comparable series exists before anyone asks for one. A record generating a margin estimate at the moment of quotation follows, and that record is kept when the proposal is made rather than when it is approved, so that the assumptions behind a pricing decision are fixed before the outcome becomes visible and cannot be reconstructed favourably afterwards.
The second stage consists of putting the variance to work. A monthly margin bridge, decomposed into price effect, cost effect, mix effect, and volume effect, is read against a standing agenda, and it becomes functional only to the extent that it ceases to be a reporting ritual and turns into an input to the price list, the supplier negotiation, and the thresholds applied to quotations. Below-threshold pricing authority is distributed by role, exceptions are logged, and over time that log becomes one of the documents a reviewing party reads most attentively — because what demonstrates the terms on which a company has defended its margin is not the average ratio but the record of the work it declined.
What a buyer is ultimately looking for in gross margin is not a high number. It is a margin that means the same thing across periods because it is computed the same way, that is produced through the same logic irrespective of who sets the price, and that signals its own deterioration before profit falls. A structure carrying those three properties remains a candidate for a higher multiple even where it sits several points below the sector average, for the straightforward reason that what is being acquired in the first case is an outcome, and in the second a capacity the company can reproduce.