At a diligence table, when the buyer's adviser asks who authorises price exceptions beyond the standard discount band, the institutional answer given in the room and the pattern legible in the records rarely coincide. The delegation of authority matrix places exception approval with the commercial director; the correspondence file, read chronologically, shows each exception passing through the founder, frequently in a single-line message and, in a meaningful share of cases, in a telephone conversation never reduced to writing. This is not a finding of procedural irregularity, and it is seldom presented as one. It is a finding that the company's operative decision architecture differs from its documented architecture. The question that carries weight in the review is therefore not whether the exceptions were correctly priced — in most cases they were, which is precisely what makes the finding difficult to raise — but whether the judgement that produced the correct pricing resides anywhere other than in a single person.
The same pattern, once identified in one place, tends to reappear across every other surface of the business. The founder attends the annual renewal discussions with each of the three largest customers personally; the extension of supplier payment terms is obtained not through a formal written request but through a relationship built over several years and understood on both sides without documentation; a key engineer's intention to leave is resolved not within a human resources process but over a dinner, and the terms of that resolution exist nowhere in the personnel file. None of these conversations produces a minute, because none of them generates the kind of institutional friction that would require one — the counterparties know one another, the interlocutor is unambiguous, and the outcome arrives quickly. A material portion of the company's operational performance is generated by exactly this speed, which is why nobody inside the organisation has ever had occasion to name the arrangement as a problem.
The structure has a name — founder dependency, the concentration of critical relationships, pricing judgement and institutional memory in one individual — and it is better understood not as an error in reasoning but as a shortcut that genuinely lowers cost under a specific set of conditions. At small scale, when information is held in one head, coordination cost approaches zero: where the person who knows which line item the counterparty will concede during negotiation is also the person who signs off on the concession, no intermediate layer, briefing cycle or internal alignment round is required, and the decision arrives inside the window in which it still has commercial value. Trust substitutes for process, and at that stage the substitution is demonstrably cheaper. The difficulty lies not in the shortcut itself but in its persistence after the conditions that made it rational have quietly ceased to hold.
That persistence is sustained by a mechanism that feeds itself. Because the cost of delegating is measured instantaneously on each occasion, doing it personally always appears faster; delegation registers as a loss on today's calendar and as a gain on a balance sheet several years out, so the comparison is structurally biased against transfer every time it is made. Second-line managers, after a small number of attempts in which their judgement was revisited or reversed, learn to escalate — the cost of escalating being low and the risk of not escalating being asymmetric — and the learning is rational at the individual level even where it is corrosive at the institutional level. The organisational chart therefore continues to broaden while the decision architecture does not, and formal authority is distributed while earned legitimacy remains at a single node. This divergence is characteristically the last thing an organisation notices about itself, because the chart is accurate and no one examining the chart can see a defect in it.
The genuine indicators of dependency are consequently found in two places other than the chart: the length of the approval queue, and the speed with which interruptions to the founder's calendar propagate into the operation. During a period in which the founder is unreachable for a week, the set of decisions that waited, the proposals that were not issued, the supplier discussion that was postponed and the recruitment offer that lapsed constitute the plainest available measure of the condition. That measure is rarely captured internally, because the delay does not usually manifest as a crisis; it manifests as a delay, and delays are absorbed rather than recorded. Yet the opportunity foregone in each such interval scales with the size of the business, so the same one-week absence carries a materially different cost at fifty million in revenue than it did at five. The ceiling that emerges here is neither technical nor financial; it is entirely a ceiling on decision throughput.
The most visible form of the institutional cost surfaces at a transaction table. A buyer understands that what is being acquired is not historical cash flow but the repeatability of that cash flow independent of the founder, and therefore tends to price the dependency not by reducing headline value — which invites a negotiation the seller can win — but by altering the structure of the consideration, which invites a negotiation the seller usually loses. The earn-out period lengthens and is tied to a broader set of operational metrics rather than a single revenue measure; the escrow percentage rises and the release schedule extends; the transition services agreement widens in scope and moves from twelve months to a multi-year commitment with defined availability obligations; the non-compete hardens in both duration and geographic reach; and the representations and warranties acquire a dedicated heading addressing the continuity of customer relationships post-closing. Each of these items is the same structural observation translated into contractual language.
The identical concentration produces a constraint on the financing side as well. Key-person provisions in credit agreements make the founder's departure, or a reduction of the shareholding below a specified threshold, an event of default or a prepayment trigger; key-person life insurance, assigned to the lender, becomes a cost line as concrete as a covenant restricting distributions. Where customer concentration and relationship concentration converge in the same individual, the two risks do not add, they compound: if the top three customers are carried by one person's relationships, the revenue at risk in that person's absence exceeds what the concentration table alone would suggest, since the loss is not distributed across the remaining base but correlated with it. Credit committees typically price this combination jointly rather than treating the two disclosures as independent, which is why a company can present acceptable concentration metrics and still encounter unexpected structural conditions in the term sheet.
Even in a company with no transaction on the horizon, the cost is fully operative; it simply accumulates in a different account. Senior managers who leave generally cite compensation in the exit conversation, but the operative reason is more often the narrowness of the decision space and the frequency with which decisions taken within it are overturned. That turnover accumulates not as recruitment expense, which is measured, but as customer relationships that must be rebuilt from the beginning and institutional memory that departs without ever having been written down, neither of which is measured. Pricing discipline erodes along the same path: when exceptions are granted on the judgement of a single person, the aggregate margin effect of those exceptions is never consolidated into any report, because no report is designed to receive them. At year end, the contraction in gross margin is investigated on the cost side, while its actual source is a set of undocumented approvals.
What neutralises this tendency is not individual awareness or a stated intention to delegate, but an architecture with four separable components. The first is recording the decision at the moment of proposal rather than the moment of approval: when the person preparing a recommendation commits the rationale, the underlying assumptions and the rejected alternative to writing, the founder's correction ceases to be an intervention and becomes a teachable standard that a third party can subsequently apply. The second is constructing the authority matrix on monetary and risk thresholds rather than on titles, with every exception above the threshold consolidated, together with its justification, into a single register reviewed on a fixed cycle. The third is operating relationship handover as a calendar rather than an intention — documenting, for each of the top three customers, the top three suppliers and each financing counterparty, who holds the second signature, and fixing in advance the date on which that person attends the meeting alone. The fourth is writing the pricing and exception logic not as it exists in the founder's head but as a decision tree a third party can execute without further instruction.
The only valid test of whether this architecture functions is the founder's systematic absence. A pre-announced period during which the founder participates in no decision of a defined class, followed by a structured review of the decisions taken in that window against three dimensions — delay, deviation from the established standard, and realised outcome — converts the level of transfer from a statement of intent into a measurable indicator that can be tracked across periods. Operated as a single exercise it produces little; operated as a recurring rhythm, over several cycles, it changes the incentive facing the second line, whose reflex to escalate is gradually replaced by a reflex to document its own reasoning, since documentation is what the review examines. Transfer, by definition, is measured by the quality of decisions produced in the room the founder is not in.
BEIREK approaches this work not as a culture programme but as the reconstruction of a decision architecture: mapping the approval flow as it actually operates, using correspondence and meeting records rather than the published matrix; recalibrating authority thresholds along monetary and risk dimensions; templating the decision record so that it is captured at proposal rather than at approval; assigning the second signature across critical customer, supplier and financing relationships and fixing the handover calendar against dates rather than intentions; and instituting a monthly review in which threshold breaches, delays and transferred conversations are tracked in a single register. The interval in which this work is meaningful is not the moment the data room opens but twelve to eighteen months before it, since the evidence a buyer credits is not a representation or a recently drafted procedure but a record series extending across several reporting cycles.
What determines a company's valuation is, in a substantial number of cases, not performance itself but the demonstrable capacity of that performance to be repeated independently of the founder — and that demonstration is not a document that can be produced when a transaction appears, but a record accumulated over years in which nothing appeared to depend on it. Founder dependency is therefore not a question of personality but a question of timing: the conditions under which it is cheap are entirely legible in advance, while the threshold past which it becomes expensive is typically identified only after it has been crossed, at a table where the counterparty has already priced it.