The running order of a board meeting is rarely varied: revenue, margin and cash first; operating indicators next; and last, if time permits, human capital, customer relationships and the regulatory environment. The sequence presents itself as an innocuous scheduling convention, yet items reached at the end of an agenda convert into decisions at a markedly lower rate than those reached at the beginning, because the argumentative energy of a room is largely spent within the first half hour and the remaining headings drift into informational status. Within the same meeting, an objection expressed through a numerical indicator and an objection grounded in operating experience meet unequal resistance — the first requires counter-evidence to defeat, whereas the second is commonly deferred with the formulation that the matter should be measured before it is discussed. That measurement is seldom performed, and the subject does not return.
A comparable pattern is observable in budget negotiations. A capital request whose payback period can be computed arrives with a standardised defensive vocabulary, while the outage exposure created by deferred maintenance, the probability that a key technical employee departs, or the effect on negotiating position of relying on a single-source supplier cannot be arithmetically expressed in the same terms, leaving the requesting manager without a defensible figure. Resources move toward whatever carries a number. This is not a malicious or incompetent allocation; it is the direct output of the institution's own decision procedure, and for as long as the procedure operates in that form the outcome will repeat in a predictable manner.
The mechanism underlying the pattern is identified in the management literature as the McNamara fallacy — the progressive treatment of an unmeasured quantity as first secondary, then immaterial, and finally non-existent. It advances in four steps: measurable quantities are measured and admitted to the decision; approximate proxies are assigned to those that resist direct measurement; quantities for which no proxy can be constructed are deemed negligible; and in the final step the unmeasured is assumed not to exist. What matters is that the progression occurs without deliberate choice, driven instead by the expansion logic of the measurement infrastructure itself. No institution elects to degrade its own judgement; it merely shifts its attention toward the surface on which data can be generated.
Understanding why the shortcut takes hold requires acknowledging that it is genuinely functional. Measurement-based decision-making solves two serious problems at institutional scale: it constrains arbitrariness and it makes accountability possible. Once an investment rationale is anchored to a figure, the decision detaches from the individual who made it, becomes auditable, and remains open to interrogation years later. This is the fundamental gain in the transition from a structure governed by founder intuition to one governed by institutional process, and it is not something an organisation can reasonably surrender. The difficulty lies not in the shortcut itself but in the unnoticed expansion of its scope: measurement is established as a filter for selecting which quantities enter the decision, and gradually becomes the frame that defines the boundary of reality.
That expansion accelerates as the indicator count rises. Intuition suggests that a richer instrument set should narrow the blind spot; the observed behaviour runs the other way. The more comprehensive a dashboard appears, the heavier the burden of proof falls on anyone raising a quantity absent from it, since the common language of the discussion has become wholly quantitative and a non-quantitative objection reads as conviction without evidence. Extensive measurement systems thereby construct a closed loop that filters out the information most capable of correcting them, and institutional confidence grows less because the measured domain widens than because the unmeasured domain is forgotten.
The institutional cost appears first in rework and schedule slippage. Progress on a project calendar is measured through completed work quantities, whereas the variables that actually govern the calendar are frequently the quality of the institutional relationship with the permitting authority, the transfer speed of information between the site team and the design team, and the true priority a supplier assigns the order within its production sequence. None of these corresponds to a cell in the weekly report, so the report remains green while delay accumulates quietly, and the deviation becomes visible only once a milestone has been missed. At that point the cost is not the deviation itself but the narrowed intervention window created by its late detection.
The second and more expensive cost materialises at the moment the company seeks a change of ownership or external financing. Questions asked at the diligence table routinely target quantities that never appeared on the company's own dashboard: how many customers the revenue depends upon and under what contract tenor, whether critical processes run on documentation or on the recollection of particular individuals, where pricing authority actually resides, and whether the supply chain carries single-source dependency. Such items were not managed because they were not measured, and were not documented because they were not managed; the result is that valuation is pulled downward through an earn-out structure, an expanded scope of representations and warranties, or an elevated escrow ratio, even where profitability indicators are strong. What determines valuation is not performance itself but the demonstrable claim that performance is reproducible independently of the founder and of specific individuals.
The third cost arises where measurement pressure begins shaping behaviour. When a forced proxy is attached to an unmeasurable quantity — visit frequency standing in for customer relationship depth, audit count for quality, recorded incident rate for safety culture — the institution ceases to manage the quantity and begins managing the proxy. Visit frequency can rise while the substance of the relationship thins; recorded incident frequency can fall not because the site has become safer but because the willingness to report has declined. The second case is particularly hazardous, since the indicator improves while the underlying condition deteriorates, and the institution receives a confidence signal from its own measurement system pointing in the wrong direction.
Neutralisation comes from altering the structure of the decision procedure rather than from generating further indicators. The workable intervention separates into three components. The first is a mandatory field in every decision file in which unmeasured exposures are explicitly named — a field that cannot be left blank, cannot be answered with 'not applicable', and without at least one named exposure prevents the file from proceeding to approval. The second is the attachment of every named exposure to a role, since a risk without an owner disappears as rapidly as a risk without a metric. The third is the translation of off-measurement exposures into conditions rather than quantities: answering the question of which observation would reveal the risk first, should it materialise, produces a more usable early warning than any numerical estimate of its magnitude.
The method BEIREK operates across capital-intensive projects institutionalises precisely this distinction. On every project under our management, a separate record is maintained alongside the quantitative progress report; within it, the exposures that fall outside measurement yet govern schedule and cost — the state of the relationship with the permitting authority, the actual priority position held with a critical supplier, the decision latency between design and site, the continuity risk attached to key personnel — are named, assigned to a role, and updated in each review cycle. No attempt is made to convert the record into an indicator set; its purpose is not to measure but to make the disappearance of these subjects from the agenda structurally impossible.
The second line of intervention converts counter-argument into a defined institutional role at decision moments. In the pre-mortem sessions we run ahead of FID, ahead of closing, and ahead of major contract approval, the project is assumed to have failed and the stated causes are required to originate from the set that does not appear on the dashboard; contesting an assumption inside the financial model is expressly outside the scope of the session, because that assumption has already passed through the model's own review loop. The distinction prevents the session from collapsing into a repetition of existing analysis. Findings are written into the decision record so that they remain traceable in the following cycle, since a pre-mortem without follow-through yields the same information as a pre-mortem never held.
An institution's measurement maturity is indicated not by the number of instruments on its dashboard but by its capacity to keep the quantities absent from that dashboard on the agenda. It holds that what is measured can be managed; the converse proposition — that what cannot be measured is immaterial — has never held, and the function of decision architecture is to preserve the distance between these two statements in a systematic way. The question worth asking of the next investment committee file is not which indicator has cleared its threshold, but which exposure the file passed over without discussion.