The fastest way to establish where decision authority genuinely sits in a company is not to read the organisational chart but to observe whose telephone rings before a mid-sized procurement decision is approved. The delegation instrument will state that sourcing decisions below a defined amount fall within the remit of the operations director; the same director, however, will not proceed on a decision falling well beneath that threshold without first sending the founder a short message. The message is rarely framed as a request for approval — its register is closer to notification — yet the signature is withheld until a reply arrives. The pattern repeats with sufficient regularity to constitute a structure rather than an incident, and it is almost never named internally as a problem, because the system functions, decisions are made, and the delay seldom exceeds a few hours.
A second expression of the same pattern surfaces in how decisions taken at management meetings are recorded. Minutes are kept, the resolution is written out, a responsible name and a date are assigned; nowhere in the note, however, does it appear against which alternatives, on which data, and under which threshold rule the decision was reached. The record preserves the outcome of the decision while discarding its reasoning. When the same item returns to the agenda a year later, the rationale for the earlier decision resides not in the institution but in the recollection of the three people who happened to be in that room, and once one of those three leaves the company, a portion of that rationale is lost permanently, leaving the successor to reconstruct from inference what was once known with certainty.
The mechanism underlying this behaviour is not a management failing but a residue of the transition in scale. In a company's earliest period the locus of decision and the locus of information coincide in a single person; the founder's approval functions less as an exercise of authority than as a verification of information, that person being the only one carrying the entire context. Under those conditions centralised approval is fast and genuinely reduces the cost of error, which makes it rational rather than deficient. The difficulty lies not in the shortcut itself but in its persistence once the conditions change: as the company passes into a size at which context can no longer be held by one person, central approval ceases to be verification and becomes a constriction in the decision line — while the habit generates no signal announcing that it has lost its justification.
A second mechanism is the tendency to define delegation almost exclusively by reference to a monetary threshold. Authority matrices are constructed along the axis of financial magnitude — up to one figure the unit manager, up to another the general manager, above that the board. A substantial portion of institutional risk, however, accumulates in decisions that magnitude does not measure: a shift to a single-source supplier, an undertaking given to a customer outside the contract, the relaxation of a technical standard, an amendment to the terms of a key employee. The monetary consequence of such decisions is small at the moment of signature, whereas their tail extends across an entire budget cycle. Where the amount threshold is used as the sole dimension, the most expensive decisions become capable of being taken at the lowest level of authority without triggering any control mechanism.
At the diligence table this structure presents itself not as an institutional question but as a very plain verification request. The reviewing party reads the authority matrix, then asks for the files of several decisions taken over the preceding twelve months, and examines one thing only: whether the approval chain contains the individuals the matrix contemplates, or whether it carries a signature or a trace of correspondence from someone the matrix does not name. That comparison is precisely where the distance between existence and practice is measured. The presence of the document produces no finding on its own; the inconsistency between the document and the file produces one directly, and in the report it is typically raised not under organisational risk but under weakness in the control environment, a heading that tends to attract broader scrutiny.
Measurement is the layer of decision hierarchy that is built least often. The large majority of companies have defined indicators for sales, production, and collection cycles while keeping no indicator whatsoever for the decision process itself, even though what could be measured here is simple and derivable from records that already exist: the elapsed time between an approval request entering the agenda and its resolution; the proportion of decisions above a given threshold in which alternatives were formally evaluated beforehand; the number of files in which a signature outside the matrix appears; and the ratio of decisions in whose approval chain the founder appears to the total number of decisions taken. The trajectory of that last ratio over time is the only objective evidence of an institutionalisation claim; where the curve runs flat, the number of layers added to the organisational chart carries little analytical weight.
The channel through which the shortfall reaches valuation is generally not, as is often assumed, a direct discount to the multiple. Where a buyer finds the decision hierarchy weak, the typical behaviour is to reconstruct the payment structure before touching the headline figure: a portion of the consideration is tied to an earn-out, the earn-out period is paired with a service condition requiring the founder to remain, a schedule of decision types requiring buyer consent after closing is added to the agreement, and a separate warranty is opened confirming that no transaction has been executed outside the scope of delegated authority. Each of these provisions lengthens the seller's path to cash and narrows post-closing freedom of movement; their combined economic effect is typically larger than a straightforward reduction in price, and considerably harder to claw back in negotiation.
The second channel becomes visible on the credit side. In a financing structure the covenant package is not built on financial ratios alone; the negative undertakings, which restrict the borrower from taking specified categories of decision without the lender's consent, form the body of the package. Where decision rights within the company are not defined in a documented and traceable manner, the lender writes those undertakings more broadly and sets their thresholds lower, substituting its own control for the internal control it cannot verify. The result is an operational flexibility narrowed by contract, and the cost of that narrowing does not accumulate in the interest margin — it accumulates in the decisions that cannot be taken over the following three years, each requiring a consent process that the counterparty has no commercial incentive to expedite.
The first component of a structural intervention is to lift the authority matrix out of a single dimension. In an architecture that functions, authority is defined across three axes together: monetary magnitude, reversibility, and the duration of the commitment created. Decisions that are irreversible or that generate an obligation extending beyond a year are escalated to a higher approval layer irrespective of amount; conversely, decisions that are large in value but reversible and recurring are pushed downward. The second component is that the decision record be kept at the moment of proposal rather than at the moment of approval — the person bringing the proposal records against which alternatives, on which data, and under which assumption the decision is being recommended, with the approval landing on top of that record. The third component is that, above a defined threshold, the counter-argument role be assigned to a named individual; that role constitutes an obligation rather than an opinion, and it is minuted.
BEIREK's intervention in this area does not begin with the drafting of a new policy; it begins with a retrospective map of the decisions already taken. The decision files of the preceding twelve to twenty-four months are reviewed, the actual approval chain of each decision is reconstructed and compared against the written matrix, and the resulting deviation table shows — at the level of decision type rather than of individuals — where authority has been delegated on paper but retained in practice. On that table the three-axis matrix is rebuilt, the decision record template is moved to the point of proposal, and a monthly review rhythm is instituted in which a single question is asked: how many files in the past month carried a signature outside the matrix, and for what reason. The trajectory of that deviation count across successive months becomes the indicator of institutionalisation itself.
The continuity dimension only becomes testable once this chain of records has begun to accumulate. What a reviewing party seeks as evidence of continuity is neither a statement of intent nor a draft succession plan; it is the files of real decisions above the threshold in which the founder does not appear in the approval chain. Where those files exist, the assertion that the decision hierarchy operates independently of the founder rests on a verifiable footing, and the company's performance is read not as a contingent outcome attached to an individual but as an institutional capacity capable of being reproduced. Where they do not exist, the assertion — however coherently it is articulated across management interviews — remains unverified in the diligence report, and every unverified structure finds its counterpart in the pricing model as a risk premium.
The difficulty of the decision hierarchy question, from the owner's perspective, is that the deficiency causes no pain in daily operations: decisions continue to be taken, the business continues to run, and no one raises the allocation of authority as a complaint. The structure becomes visible only when an external party sets out to verify it, or when the founder is not at the table, and both circumstances usually arrive after the moment at which the structure ought to have been built. The practical criterion for when a company should construct its decision architecture is therefore not a threshold of size: at the point at which the proportion of decisions carrying the founder in the approval chain begins to exceed the proportion of decisions whose context the founder genuinely commands, the architecture should already have been in place.