By the second week of an investment review, when the organisational chart sitting in the data room is placed alongside the notes from that same week's management interviews, a recurring pattern usually emerges: a meaningful share of the decisions distributed across three separate functions on the chart are closed in the interviews with a single sentence — that the item was settled in a conversation with the general manager. The chart itself may be faultless, with hierarchically consistent boxes, clean reporting lines and titles allocated in a reasonable manner. Yet when the supplier switches, pricing exceptions, hiring approvals and capital expenditure decisions of the preceding twelve months are opened one by one, the signature that actually closed the decision belongs, in most instances, not to the person who appears to own that item on the chart but to the level above. This is not a discipline problem; it is the natural consequence of a company whose growth rate has outpaced the rate at which its delegation mechanics were built.
Internally the situation is rarely named, because it is functional for everyone involved. A middle manager who escalates a decision is not buying decision speed but shedding decision risk; a founder or general manager who pulls the decision back retains a felt sense of control over the quality of the outcome, and often does so on defensible grounds, given that a substantial portion of the institutional memory genuinely resides in that person's head. Up to a certain scale the trade is rational: while decision volume remains low, the cost of routing a decision to whoever holds the most context is lower than the cost of building a delegation architecture. The difficulty lies not in the shortcut itself but in its persistence once the underlying condition changes — once monthly decision volume rises several-fold, or once geographies and business lines multiply.
The RACI matrix is precisely the instrument expected to enter at this point, and precisely the instrument most frequently misapplied at this point. Its four roles — the party executing the decision (Responsible), the party answering for the outcome (Accountable), the party whose view must be taken before the decision is closed (Consulted), and the party informed afterwards (Informed) — differ in nature, the first two governing authority and the latter two governing information flow. The most commonly observed degradation is the assignment of the Accountable role to more than one individual, or to a committee. Accountability does not strengthen when divided; it evaporates. Where two people share responsibility for a decision, the moment a disagreement arises no mechanism remains other than escalating the matter upward, and the matrix — drafted to reduce founder dependency — becomes the document that institutionalises it.
A second degradation appears where the Consulted role is left empty. If the parties whose views must be obtained before a decision can be taken are not written down, the obligation to consult does not disappear; it merely becomes unpredictable. Operationally, this means part of the process stalls in legal, part in finance, and part nowhere identifiable at all; and because the source of the delay is undocumented, an extended cycle is read inside the organisation not as a process defect but as the individual slowness of the relevant manager. That misattribution then routes the corrective effort to the wrong place — the person is replaced and the cycle time does not shorten.
What the reviewing party is looking for, accordingly, is not the existence of the matrix. Finding a signed and dated RACI document in the data room closes the first question and opens the substantive one: whether recent decisions actually travelled the paths the matrix prescribes. That question is testable retrospectively, and an experienced review team tests it not by reading documents but by drawing a sample — a random selection from expenditure approvals above a defined threshold, non-standard pricing exceptions, contract amendments and key-personnel hires, with the decision chain of each traced backwards. To the extent the signature at the end of the chain diverges from the name recorded in the matrix, the evidentiary weight of the document declines accordingly.
The measurement dimension is the layer most often skipped, largely because few companies consider a RACI matrix to have a measurable output at all. The health of the matrix can nonetheless be tracked through at least three indirect indicators: the distribution of decision cycle times for above-threshold items, the proportion of decisions escalated beyond the owner named in the matrix, and the frequency with which a decision is reopened — that is, returned to negotiation after having been closed. The second of these measures the de facto validity of the delegation architecture directly; the third captures the absence of a defined Consulted set, since a late-arriving view reopens a settled decision. None of these indicators requires a sophisticated system; the approval workflow already generates them, and the only missing step is a habit of looking back at what it produced.
The way this deficiency reaches the valuation is, contrary to common assumption, seldom through the multiple. An acquirer facing a company whose delegation architecture is documented but not practised tends to frame the exposure in these terms: once the founder steps away, a portion of decisions will simply not be taken in the near term while another portion will be taken at the wrong level, and the cost of that concentrates in the first four to six quarters after closing. Such risk is typically managed structurally rather than through price — a service agreement binding the founder through a transition period, an earn-out tranche linked to post-closing performance, an expanded representation and warranty perimeter covering out-of-ordinary-course decisions, and, characteristically, an elevated escrow ratio. Each of these items pushes the timing of the seller's cash flows outward; the aggregate effect shows up not in the headline figure but in the present value of what the seller actually receives.
The same mechanism operates on the credit side. An undocumented delegation architecture does not, in itself, generate a dedicated covenant heading in a financing structure; it does, however, tend to lead a credit committee to tighten information and reporting covenants, to impose prior-consent requirements for defined transaction types, and to draft the key-person clause more narrowly. The resulting cost is less an increase in the price of debt than a loss of post-closing operational latitude — and that loss is of the kind that grows more expensive as the business scales.
Structural intervention does not begin with rewriting the matrix; it begins with separating out which decisions warrant one at all. In most companies the overwhelming majority of decisions taken are repetitive and uncontested, and migrating them into a matrix renders the document unusable. A meaningful separation is generally made across four headings: commitments exceeding a monetary threshold, commercial terms departing from standard, technical and procurement decisions that are costly to reverse, and decisions concerning key personnel. For everything outside these four groups a simple authority limit suffices rather than a matrix; the weight of a RACI structure earns its keep only within them.
In complex, capital-intensive projects, BEIREK builds this separation from the decision record rather than from the management chart. On the engagements the firm runs, every above-threshold decision is attached to a one-page record before it is taken: who executes it, who answers for the outcome, which views cannot be bypassed before it is closed, and on what date and on what information it was made. The record is opened at the moment of proposal, not at the moment of approval, because a record kept at approval documents only the outcome, whereas a record kept at proposal reveals whether the delegation architecture is actually functioning. On a monthly rhythm, three readings are taken from these records — whether the decision remained with the owner named in the matrix, the proportion of decisions escalated, and the consultation link at which delay accumulates — and the matrix is then revised on the findings of that reading rather than through a theoretical redesign exercise.
The measure of continuity emerges at the same point. What indicates whether a RACI matrix has matured into an institutional capability is not when it was written but how many times, and on what trigger, it has been revised. A matrix drafted on a single date and untouched thereafter tends to be read on the reviewing side as an advisory deliverable, with its evidentiary weight discounted accordingly; a matrix updated at each new business line, each new geography or each threshold change, with the rationale for the update recorded, indicates instead that the company carries delegation as a working governance function rather than a one-off document. The difference between the two situations is that the same artefact belongs to two entirely different classes of evidence.
The question that reveals the most about a company's authority architecture is not who approves what, but which decisions are placed on hold when the founder is unreachable for two weeks. No organisational chart records that answer, yet any company keeping a decision record can produce it within fifteen minutes; and what determines the valuation in a review process is, more often than not, the company's demonstrated capacity to produce that answer at all.