Set two decks from the same company side by side, three months apart, and a pattern surfaces often enough at investment committee to be worth naming: the total addressable market figure has grown appreciably between them, while the customer count, the contracted volume and the number of live opportunities in the pipeline have barely moved. Asked to account for the increase, the sponsor typically offers not a demand observation but a definitional revision — an adjacent user segment folded into scope, a geographic boundary extended by one more country, a budget line sitting beside the spend the product displaces added into the total. The number is larger. What has grown, however, is not the market but the frame within which the market was drawn.
A second face of the same pattern becomes visible when attention shifts from how the figure was produced to which threshold it was produced to clear. Where a fund's minimum return expectation, a holding company's investment hurdle or a board's growth mandate establishes in advance the floor a project must reach in order to remain defensible, the market estimate tends to be constructed backward from that floor; the arithmetic is performed, but it runs in reverse. None of this requires anyone to distort data. Because the category boundary is elastic, arriving at the required magnitude demands no bad faith whatsoever — and where no written rule governs where that boundary sits, every drawing of it appears equally defensible.
The behaviour has a name: TAM inflation, the enlargement of total addressable market through definitional widening rather than through observed growth in demand. Its mechanics rest on two components. The first is the structural elasticity of top-down arithmetic — a broad category total multiplied by an assumed share can shift by an order of magnitude on a modest adjustment to either variable, and the smallness of the multiplier, a share assumption of one percent or less, lends the calculation an air of conservatism it has not earned. The second is the anchoring effect of the first figure spoken aloud: whatever magnitude reaches the table first becomes the reference around which all subsequent discussion turns, and the negotiation proceeds not over whether that reference is correct but over what discount should be applied to it.
This tendency deserves treatment not as an error but as a shortcut that is functional under specific conditions. At a stage where the product has not yet been sold, conversion has not been observed and pricing remains untested, a top-down estimate of magnitude is not a measurement but a coordination device, aligning through a shared and openly provisional premise the scale at which teams think, the architecture they build and the speed at which they organise. Precision is not what anyone expects of the number at that point. The difficulty lies not in the shortcut itself but in its persistence once conditions change — once the first genuine sales cycles have closed and win rates, cycle lengths and price elasticity are in hand, a figure built for coordination continues to stand in for measurement.
What sustains that persistence at the institutional level is that the number belongs to no one. The revenue budget has an owner, cost lines have owners, the cash projection sits with the finance director; the addressable market figure, in most organisations, appears in nobody's performance review. A number without an author acquires no falsifier — nobody defends it, so nobody contests it, and it settles into institutional memory as a datum rather than as an assumption. Several planning cycles later, no one remains who recalls the premises under which the boundary was first drawn, yet the figure continues to travel from deck to deck, gaining apparent authority with each repetition.
The cost materialises first not in valuation but in the hiring plan. Sales headcount is typically derived from the addressable market: the number of accounts the market is held to contain, divided by the accounts a representative can carry, multiplied by a target coverage ratio, yields headcount; headcount yields quota, quota yields the revenue target, and the revenue target yields the burn profile. A definitional widening at the head of that chain becomes fixed personnel expense some three quarters later — and personnel expense is among the least reversible lines on the statement when revenue arrives late, since a reduction in force affects the pipeline and the firm's standing in the labour market at the same time. The cost of inflation therefore arises less from the estimate being wrong than from its having been converted into commitments that cannot be unwound.
In capital-intensive projects the same mechanism appears on a harder surface, because there the definitional widening is converted into capacity rather than into headcount. The scale of a production line, a plant or a data centre campus is set as a function of projected demand; once scale is fixed, land, interconnection capacity, long-lead procurement and frequently take-or-pay supply contracts are committed against it. Where demand fails to present itself at the projected volume, the shortfall surfaces not as a marketing problem but as fixed cost per unit: utilisation falls, unit cost rises, the rise in unit cost narrows the room for price competition, and the loop tends to become legible on the testing date of a DSCR or leverage covenant. What the lender observes at that point is not a forecasting error but a debt service gap.
At the transaction table the cost takes another form. Buy-side diligence teams rarely test a market estimate on the terms in which it is presented; they redefine the unit of demand and rebuild the figure bottom-up, beginning with who makes the purchasing decision, out of which budget line it is funded and at what approval threshold it clears. The gap between the two numbers is generally settled not through a headline price reduction but through mechanisms embedded in the structure of consideration — a portion of the purchase price deferred into an earn-out tied to revenue thresholds, a portion held in escrow, with representations and warranties tightened around customer concentration and contract renewal rates. The more durable effect is indirect: an estimate that cannot be reconstructed bottom-up invites the same treatment of every other assumption, and that widening inquiry lengthens the timetable while expanding the list of conditions precedent.
Neutralising the tendency calls for institutional architecture rather than an appeal to individual restraint, and four components carry most of the weight. The first is a written definition of the demand unit — whether the thing being counted is a company, a facility, a line or a contract, and which spending line pays for it. The second is construction of the estimate backward from the purchasing decision: who decides, what approval threshold applies and how long one cycle runs. The third is freezing the definition, so that once the category boundary is drawn, every subsequent change to it is recorded as a discrete decision and growth in the figure is reported split between the portion attributable to observed demand and the portion attributable to redefinition. The fourth is assignment of ownership; from the moment a number carries a name, it is obliged to be defensible.
BEIREK's intervention at this point is not to raise or lower the estimate but to install a record-keeping order that makes visible which component the estimate came from. Three registers are maintained in practice: a definition register, in which the demand unit and the category boundary are fixed with a date and a named owner; a revision register, stating the rationale for each revision and separating how much of an increase is definitional from how much is observational; and a loss register, accumulating the stated grounds on which opportunities were declined or never opened at all — since the true perimeter of an addressable market is read less from the business won than from the reasons a purchase never came up for consideration.
The second line of intervention decouples the estimate from irreversible commitments. Headcount, capacity scale and long-term supply obligations are tied not to a single market scenario but to a staged structure whose base tier is the band that observed conversion data supports; the opening of each subsequent tier is locked to a trigger defined by executed contract count and cycle duration rather than by a revision to the market estimate. Sensitivity analysis is run on the definitional variable instead of the customary growth rate, testing whether the structure survives when the category boundary is narrowed by one segment, because what usually renders a project fragile is not a deviation in growth rate but the surplus fixed cost that appears the moment the definition is drawn one step tighter.
A market estimate matters not because it is right or wrong but because of the commitments it triggers, and a firm's forecasting discipline is measured less by the number itself than by which decisions that number is permitted to lock. Where it is not documented who drew a given magnitude, under what definition and on what date, the magnitude in question is not an estimate at all; it is a commitment that entered institutional memory disguised as an assumption.