When an investment committee presentation reaches the market section, three concentric circles are typically displayed, and the outermost figure passes without objection; the objection begins at the innermost circle, at the slice the company claims it can genuinely capture. The question asked at that moment is almost always the same one — where does this percentage come from. The delay before an answer arrives reveals, ahead of the answer's substance, whether the estimate is a structure built inside the company or a number produced while the deck was being assembled. Where the delay runs long, the person who prepared the figure is the only person in the room who can defend it, and that person is more often than not the founder. This is not an arithmetic failure; it is the surface at which an institutional gap becomes visible on the presentation table.
A second observation surfaces across two documents belonging to the same company. The serviceable market figure appearing in the business plan submitted to a lender, and the prospect count underlying the table from which the sales team's annual quota is derived, are prepared under two logics unaware of one another, and they rarely reconcile. The business plan figure is built top-down, by applying a share percentage to a total drawn from an industry report; the sales table figure is built bottom-up, from meetings per representative. The gap between the two is sometimes a multiple rather than a margin, and the fact that no one has noticed the gap indicates that neither figure has ever entered the same management decision.
The mechanism underneath this behaviour is not a deficiency in forecasting skill but the fact that the forecast serves two distinct functions. A top-down market figure exists to persuade; it establishes magnitude, not boundary. A bottom-up figure exists to allocate resources — determining how many representatives are hired and how many are deployed into which territory — and for that reason it is structurally obliged to remain conservative. Two functions bring with them two owners and two levels of prudence. The difficulty lies not in the divergence of the two methods but in the fact that the company has never decided which of them is binding; while that decision remains unmade, both figures stay defensible and neither becomes accountable.
The shortcut is functional under identifiable conditions. At an early stage, before the product has settled, constructing a serviceable market upward from sales capacity is simply not possible: there is no historical win rate, average contract value rests on a handful of observations, and the length of the sales cycle remains unknown. Leaning on an industry report under those conditions is a reasonable cost-reducing shortcut. Once the company has completed a meaningful number of repeated sales cycles, however, the condition changes; it now possesses its own win rate, its own documented loss reasons and its own boundary of geographic reach. Continuing with the shortcut past that point becomes an election rather than a constraint, and the review table reads it precisely that way.
The institutional counterpart of serviceable market is not, in substance, a market question but a capacity question. The slice a company can genuinely capture over the coming twelve months is a function of how many representatives it fields, in which territories, through which dealer or distributor network, at what lead time and against what production or service capacity — not a function of the industry's aggregate size. A properly constructed estimate is the product of these variables, and each variable carries its own evidentiary source: headcount from payroll, win rate from the sales record system, average contract value from invoices, throughput from the production plan. Once that chain is assembled, the estimate ceases to be an assertion and becomes a derivable result; absent the chain, the figure remains a percentage without provenance.
The channel into valuation opens here. The reviewing party validates the first year of a revenue projection against contracted work and backlog, while the second and third years rest largely on the serviceable market assumption. Where that assumption carries no evidentiary chain, the weighted portion of the projection is treated as unverified, and the consequence typically appears in one of two forms: either the multiple is discounted directly, or — preferring to defer the risk into the payment schedule rather than extract it from price — a material portion of consideration is tied to an earn-out. The second route is selected more frequently because it preserves flexibility for the investor; from the seller's side, the result is acceptance of a performance threshold measured by a mechanism the seller did not build.
A second channel emerges within the representations and warranties package. Statements concerning market size are generally treated in the agreement as matters of opinion and are not directly indemnifiable; the underlying data feeding that statement, however — the customer list, the status of pipeline opportunities, the territories covered by dealer agreements — is treated as verifiable fact. Where a company has not derived its serviceable market from those underlying records, the data room shows those line items as either incomplete or internally inconsistent, and once inconsistency surfaces the negotiation shifts from a discussion of market share to a discussion of information quality, which is invariably the more expensive of the two. An elevated escrow percentage is frequently the residue of that second discussion.
The third channel is founder dependency, and it operates most quietly. Judgment about where the serviceable market narrows and which segment it widens into sits, in most companies, with a single individual and appears in no document; that individual knows intuitively which customer will actually buy, which tender was effectively closed before it was announced, and in which territory a competitor cannot move on price. This intuition is usually correct, and that is precisely what makes it hazardous, because as long as it remains correct no one feels the need to commit it to record. When the reviewing party observes this configuration, the question becomes whether what is being acquired is a market position or a revenue stream conditional on one person remaining, and the second answer triggers key-person undertakings and pre-closing conditions directly.
The intervention that neutralises this tendency is not heightened individual attention but an architecture composed of three separable components. The first is designation of a single binding version of the estimate, fixed as the sole source feeding the budget, the sales quota and the investor presentation alike; disallowing two figures to live in parallel eliminates, on its own, most subsequent variance disputes. The second is documentation of the estimate at the input level, with the source, measurement date and responsible party for each variable held in the record, so that when the figure moves it is visible which input moved. The third is a quarterly retrospective rhythm in which the slice forecast for the prior period is placed alongside the work actually won, with the reason for the variance recorded in writing.
BEIREK's intervention in this area typically begins not by recalculating the estimate but by reconstructing its derivation chain from outside the company: win rate extracted from sales records, average contract value from invoices, reach capacity from payroll and the territory plan, and from these an independent bottom-up figure produced. No attempt is made to close the gap between that figure and the one the company has stated; the gap itself is recorded as a finding, with each contributing assumption flagged individually. Ownership of the estimate is then attached to a single role — a role that sits, in most cases, on the commercial side rather than in finance — and the quarterly variance report produced by that role is run as a standing management rhythm. The objective is not to produce accurate forecasts; it is to make it possible, on the day variance appears, to state which assumption produced it.
The continuity dimension of this architecture is achieved not by eliminating founder judgment but by converting it into record. Once it is written down why a given customer is treated as unreachable, why a territory has been excluded from scope, and why a segment is priced conservatively, that judgment is attached to the company's record rather than to one person's presence; if that person departs, what suffers is not the judgment itself but only the speed at which it is refreshed. This is precisely the distinction an investor is looking for, and it is usually expressed in review reports in a single sentence: the market estimate is reproducible by the company. The ability to write that sentence is independent of whether the estimate proved correct, and its effect on valuation is frequently the larger of the two.
The party examining a serviceable market estimate at the review table is not, in truth, examining the market; it is examining how honestly the company can draw its own boundaries. A company that draws that boundary narrowly and with stated reasons is assessed at a materially higher level of confidence than one that draws it broadly and without them, because the first has measured its own capacity while the second has merely restated the size of its industry. The question worth asking is not how large the estimate is, but whether the company could reproduce it from the records in front of it, with the founder out of the room.