In a board meeting reviewing a performance pack that has carried the same format for three consecutive years, nearly the whole of the discussion is allocated to variance: which region fell short of target, which line item exceeded budget, which receivable slipped past terms. A session in which the pack itself is interrogated — which metrics are tracked, and which business-model assumption each of those metrics was originally designed to test — does not occur for years in most companies. This is not inattention; attention is in fact concentrated exactly where the institution has taught itself to look. The difficulty is that the market conditions prevailing when the measurement set was designed may no longer hold, and that lapse is structurally invisible from within the same measurement set.
A sharper indicator, available in the same meeting, is the vintage of the customer anecdotes offered by the founder or the chief executive. The narrative describing who the company is, why it gets chosen, and how it separates itself from competitors is usually built on observations gathered during the period of fastest growth, a period that may sit five or, in some cases, ten years in the past. That narrative is not false; were it false, someone would have noticed. It is simply that the market it describes has ceased to exist in the form described. The sales organization encounters this weekly in the field, but the channel through which field experience travels upward is the variance report, and a variance report communicates that a target was missed, never that an assumption has lapsed.
The name for this pattern is strategic drift — the gradual loss of external validity in a business model that continues, throughout, to preserve its internal consistency. The mechanism operates as follows. Each adaptive decision is correct at the moment it is taken and correct against the measurement set then in force. Under margin pressure the scope of a service line is narrowed, because that line is not profitable; proposal preparation time is compressed, because conversion rates permit it; a particular customer segment is exited, because its collection performance is weak. Taken individually, every one of these decisions is defensible, and data can be produced in its defense. Taken together, they relocate the company to a market position that no one consciously selected.
Ignoring the conditions under which this tendency is functional would weaken the diagnosis. Gradual adaptation is the ordinary form of organizational learning; a company that reconstructed its strategy at every signal would destroy its institutional memory and render execution discipline impossible. Commitment to the existing model, for as long as the conditions underpinning that model hold, preserves the cumulative return embedded in the learning curve, in supplier relationships, and in operational routine, and these are real and expensive assets. The problem lies not in the shortcut itself but in the absence of any mechanism that checks whether the condition validating the shortcut still obtains. Erosion arises not because commitment is wrong, but because commitment becomes unconditional.
The first place the institutional cost appears is not profitability. Profitability can be defended for a considerable period, since in an eroding model the cost side is tightened and the margin is held for another several cycles. The cost surfaces first in pricing elasticity: the rate at which the same proposal is accepted at the same price declines, the point at which a discount request arrives shifts toward the end of the negotiation, and the sales cycle lengthens at a pace slow enough to escape notice. Customer acquisition cost then rises, though the rise dissolves inside the aggregate marketing budget and is never interrogated as a separate line. What these three indicators share is that none of them, taken alone, crosses an alarm threshold, while together they describe a business model whose purchase in the market is narrowing.
A second layer of cost accumulates in the composition of the portfolio. An eroding model concentrates the company toward whatever customer segment still values it, and that segment is typically older, slower-growing, and burdened with a high cost of searching for alternatives. Customer concentration increases, the average age of the revenue relationship rises, and the share of total revenue contributed by customers won in the trailing twelve months falls. Placed side by side on a diligence table, these three ratios contradict the growth narrative in plain terms. Within the company's own reporting, however, they are rarely assembled in one place, because each of them lives inside the report of a different function.
In a sale or investment process, the way erosion translates into valuation is direct. The buy side asks how much of the historical performance is repeatable under present market conditions, and answering that question requires demonstrating which assumptions past growth was built upon. Where the assumptions were never written down, which is typically the case in an eroded company, the answer collapses into the founder's verbal account, and a verbal account cannot transfer risk to a counterparty. The consequence appears either as a discount applied to the multiple itself, or as a reallocation of consideration into an earn-out structure, the addition of conditions precedent, and an increased escrow percentage. What determines valuation here is not the level of performance but the demonstrability of the conditions under which it will recur.
The intervention that neutralizes erosion is constructed through institutional architecture rather than individual foresight, and it has three components. The first is an assumption register: the five to eight core assumptions carrying the business model — who buys, for which need, against which alternatives, within which price band — are written in explicit sentences, and beside each is recorded the observable signal that would indicate the assumption has lapsed. The second is an externally referenced measurement threshold: triggers defined against market behavior rather than internal targets, such as win rate, price acceptance rate, and new-customer revenue share, with breach of a threshold initiating a review independent of any budget variance. The third is a counter-argument role: within the review session, responsibility for arguing the scenario in which the present model has already lapsed is formally assigned to a named individual, whose view is recorded in writing alongside the decision.
BEIREK constructs this architecture, in capital-intensive and financed projects, through decision-record discipline. At project or portfolio level, the assumptions carrying an investment decision are written at the moment of proposal rather than at the moment of approval; each assumption is paired with the observation that would invalidate it and with the person accountable for watching for that observation, so that subsequent discussion proceeds on the question of which assumptions held and which failed, not on the outcome alone. To break the anchoring effect through which the budget cycle fixes attention on the prior year, the strategic review is operated on a separate calendar; the session in which budget lines are negotiated and the session in which the external validity of the business model is questioned are not merged into one agenda.
A second line of intervention is established on the surface where erosion is most visible, namely counterparty behavior. A systematic record is maintained of lost proposals, cancelled tenders, and the items removed from scope during negotiation, and this record is used not to measure sales performance but to read how the market prices the company's offer. On the supply side, the same discipline traces which contractual clauses have hardened across the past three years of negotiation with suppliers and contractors, since a shift in bargaining power along the value chain is frequently the first tangible signal of a shift in the position of the business model itself. Read together, these two records make erosion visible one to two budget cycles before it reaches the financial statements.
What these mechanisms share is that none of them depends on the awareness of the decision-maker. The defining property of drift is precisely that each decision producing it appeared reasonable at the moment it was taken; an intervention recommending greater vigilance therefore fails to solve the problem at the level at which the problem has been described. Structural intervention aims instead to leave, at the moment of decision, the material basis for a future interrogation: a written assumption, a defined threshold, an assigned counter-argument. Once those three are in place, the institution does not change its strategy more often; it becomes able to decide whether to change or to hold, independently of the anchor set by the prior decision.
A company rarely loses its market position through a wrong decision. Far more often it loses that position by continuing to repeat a correct decision after the condition that made it correct has disappeared. The operative question is therefore not how accurate a management team's decisions have been, but through which mechanism and on which cadence it verifies whether the assumptions underlying those decisions still hold. Where no such mechanism exists, the market selects the timing of the verification, and the timing selected by the market is, as a general matter, considerably more expensive than the timing an institution would have chosen for itself.