In an investment committee session, a slide that constructs market size from the bottom up passes with noticeably fewer objections than one that constructs the same figure from the top down. The second invites scrutiny in its opening sentence, since claiming a percentage of a total market is understood to be a claim rather than a calculation; the first, presenting target customer count, unit price, and purchase frequency side by side, reads as arithmetic. Both slides, however, produce a single output, and in both cases that output has been generated from assumptions. The only difference lies in how visibly those assumptions have been decomposed, and the visibility of the arithmetic generates the impression that each number entering it has been tested — an impression that redirects committee attention to the wrong row.
Which lines get interrogated in that session, and which pass untouched, forms a recurring pattern. Unit price is debated, because everyone holds a view on price; the definition of the addressable universe is debated, because narrowing scope signals seniority; conversion rate is sometimes debated, because a reference range circulates in most sectors. By contrast, how many sites an installation crew can actually complete in a week, how many qualified conversations fit into a representative's calendar, how many new items a distribution partner can genuinely feature in a single quarter, or what monthly throughput a permitting office sustains — these typically cannot be questioned, because they do not appear in the model as a line at all. What passes unexamined passes not because it is correct, but because it is not visible.
The mechanism warranting a name at this point is the bottom-up market-sizing error — the construction of market size from unit and channel assumptions while treating delivery capacity as implicitly unlimited. The method originated as a corrective: against the emptiness of claiming a small percentage of an enormous market, it proposed rebuilding demand from individual units upward. The correction, however, operates only on the demand side. Multiplying customer count, price, and frequency, the model produces a ceiling while treating the production, installation, sales, and service capacity required to reach that ceiling not as a constraint but as an input available on demand. What emerges is not a market size but a hypothetical ceiling computed with capacity released.
Recognizing that this shortcut is functional under specific conditions is a precondition for managing it. At the exploration stage, the cost of deciding whether an opportunity merits a second look must remain low; building a full model that internalizes capacity constraints imposes a burden that cannot be carried across ten ideas none of which have yet been eliminated. Multiplying the demand side quickly to obtain an order of magnitude is, at that stage, a rational choice. The difficulty lies not in the shortcut but in its persistence after conditions change: once the same table stops functioning as a screening device and becomes the foundation of a capital allocation decision, a hiring plan, and a credit base case, the implicit assumption inside it is no longer a convenience but a commitment.
The capacity constraint does not sit in the same place across businesses, though it sits somewhere in every one of them. In installation-heavy models the binding constraint is certified technician count and crew hours; in enterprise sales it is cycle length combined with the number of open opportunities a representative can carry simultaneously; in channel-led models it is not the distributor's shelf but the distributor's attention, since a partner's sales force genuinely features only a limited number of items in any given period; in physical infrastructure it is permitting throughput, interconnection position, or the lead time on a critical component sourced from a single supplier. The model expresses growth as a demand parameter while realization depends on a capacity parameter, and the two move at different speeds. That the distinction is absent from the model does not mean the physics of the business has disappeared; it means only that the decision was taken without accounting for that physics.
The first layer of institutional cost does not, contrary to expectation, appear in the revenue line. Revenue simply lags; the damage accumulates in the fixed cost base constructed in advance against that lagging curve. Headcount planning follows target revenue, so hiring is completed at the scale demanded by volume that never materializes; facility or warehouse capacity is leased against the same curve; and both are items reversed far more slowly than revenue moves. The variance therefore settles onto the balance sheet not as a one-period shortfall in turnover but as a structural cost surplus carried across several periods.
The second layer emerges in the working capital cycle. The volume expectation produced by the bottom-up calculation calibrates inventory policy and supplier commitments; when realized capacity fails to consume that volume, inventory turnover slows, minimum purchase undertakings given to suppliers convert into a loss of flexibility, and cash locks inside a line item that does not move. The trace of this tendency on the balance sheet is usually read not in the inventory figure itself but in the ratio that figure forms with the prior year's sales volume. Receivables deteriorate in parallel, since capacity constraints delay delivery, delayed delivery prevents invoicing, and the distance between the revenue forecast and the cash forecast widens somewhat further each quarter.
The third and frequently most expensive layer sits on the financing side. Once the same curve enters the credit base case, minimum revenue thresholds, DSCR calculations, and earn-out triggers are calibrated against it; when the capacity constraint asserts itself, what is breached is not the business plan but the covenant package, and at that point the negotiating table is no longer under the developer's control. In a sale or funding round the meaning of the variance widens further: the counterparty does not confine itself to pricing a single missed period, reading the gap between plan and outcome instead as a credibility test of the entire forecasting discipline and applying the resulting discount across the whole projection horizon. The question posed at the diligence table is typically this: what evidence exists that this company knows its own capacity accurately.
This tendency cannot be managed through individual prudence, since its source is not optimism but model architecture, and the intervention must therefore also sit in the architecture. The workable structure separates into four components. First, a mandatory capacity model maintained alongside the market model — the same revenue curve reproduced in crew hours, representative calendars, line throughput, or permitting flow, with the intersection of the two treated as the only committable curve. Second, a constraint record in which the binding constraint is named explicitly, opened at the moment of proposal rather than at the moment of approval. Third, an assigned owner for every critical assumption, since an unowned assumption is defended in no review and consequently tested in none. Fourth, recalibration tied to physical thresholds rather than calendar quarters — first installation, first week at full capacity, first complete period with the first channel partner.
BEIREK constructs this intervention on projects not by correcting a single table but by binding two models to one another. Alongside every market calculation entering a capital allocation decision sits a mirror model producing the same output from capacity units; the gap between the two figures is not closed but preserved as a line item, because that gap constitutes the project's actual risk margin and becomes invisible the moment it is reconciled away. In parallel, the binding-constraint record is opened at the proposal stage and kept live through closing, with each relaxation of the constraint documented by author, by evidence, and by date, so that a post-closing variance presents itself not as a surprise but as a traceable chain of decisions.
The second mechanism disciplines how the same curve is presented, at different statuses, to different counterparties. The target curve against which the development team works need not be identical to the base case handed to financing, and in most situations separating them is the healthier arrangement; what matters is that the rationale for the difference is written down and grounded in capacity evidence. In the processes under our management, the base case remains limited to demonstrated capacity until a physical threshold has been crossed, while the upper portion of the target curve is carried as a distinct optionality layer to which covenants, headcount plans, and inventory policy are never anchored. The distinction does not prohibit optimism; it prevents optimism from converting into irreversible commitments.
The strength of bottom-up sizing lies in building the market from a unit rather than from a percentage; its defect lies in never applying that same discipline to the supply side. How accurately a company knows its own capacity is frequently more determinative than how accurately it forecasts its market, since the former is controllable and the latter is not. The question worth asking of any market calculation is therefore not how large the number is, but which line item is exhausted first on the path to that number, and where in the model that fact is recorded.