In a quarterly review, the connection between the performance packs presented by manufacturing and procurement and the three strategic priorities the board approved ten months earlier is typically established verbally, in the room; the packs themselves do not carry it. The metrics are internally coherent, most targets have been met, and unit heads defend a good quarter in their own domains with entirely legitimate grounds, while across the enterprise two of the strategic priorities have not moved at all. Producing this picture requires no bad faith, no competence gap, and no neglect; it is sufficient that no one recorded what the objective became, and at which point, as it travelled downward. When the meeting closes, everyone is right on their own measure, and the company is still not where it intended to be.
The same pattern appears in a capital-intensive project within a narrower time window and at a considerably harder cost. The return the sponsor committed to its investment committee is translated into a budget discipline at the project company, into a schedule metric at the site, and into a weekly fabrication count at the subcontractor; each link in that chain is defensible on its own terms, yet the translation assumptions sitting between the links are written down nowhere. When the site team resequences work to protect its fabrication count, what that resequencing does to the commissioning calendar — and therefore to first-draw conditions — becomes visible only once the calendar has already slipped. The metric reads green, the project reads red, and no one holds the relationship between the two.
The name for this disconnect and the remedy for it come from the same place: hoshin kanri — policy deployment — a discipline that moves strategic priorities between layers through two-way negotiation rather than one-way distribution, and that requires explicit documentation of which objective corresponds to which metric, and which metric to which activity, at each layer. Its distinguishing feature is not that senior management pushes an objective down, but that the receiving layer can push back against it on the evidence of its own capacity, resources and constraints; this back-and-forth, known as catchball, tests the realism of the objective before commitment rather than during execution. A second feature is numerical restraint: fifteen strategic priorities means, by definition, that there is no priority at all, since resource allocation acquires meaning only through exclusion.
Deployment collapsing into a one-way exercise is not a failure but, under particular conditions, an entirely rational shortcut. Negotiation is expensive; working through feasibility with every unit consumes management attention, requires data preparation, and delays decisions, whereas writing the target once and distributing it closes the same loop in a week. In a stable demand environment, where capacity is known and supply conditions are predictable, the cost of that shortcut is genuinely low. The difficulty lies not in the shortcut itself but in its persistence after conditions change: when supplier lead times extend, when input prices move outside their historical bands, or when the company enters a new market or a new asset class, the translation assumptions that were reasonable a year earlier lapse quietly while the metric pack remains unchanged.
The second mechanism is the substitution of the metric for the objective. The moment an objective is made measurable, an obligation to perform against that measure comes into existence; and to the extent the metric is not a perfect proxy for the objective — which, in practice, it never is — the shortest path to satisfying the metric is frequently a path that does not serve the objective at all. Procurement can hold its unit-price measure at the cost of narrowing the supplier base; planning can hold line efficiency through long production runs and the inventory accumulation that follows; the site team can hold percentage completion by pulling forward items that are not on the commissioning critical path. All three are behaving correctly against their own measures, and the resultant behavior is the opposite of the strategy.
The third mechanism is deployment designed vertically while horizontal dependency remains unmapped. Alignment between layers says nothing about what units at the same layer owe one another, even though most operational outcomes are produced precisely at those lateral interfaces. When a supplier development target sits with procurement, a quality rejection target with manufacturing, and a delivery reliability target with logistics, the fact that all three cannot be met simultaneously under a given supplier configuration appears in no report. The conflict does get resolved somewhere during the year — generally in favor of the unit holding the greatest formal authority rather than the unit holding the strongest argument, which is a governance outcome rather than an operational one.
On the balance sheet, these three mechanisms rarely register as a single identifiable expense line; they register as three unconnected lines moving in the same direction: inventory turns slowing, expedited freight and emergency procurement growing as an off-budget item, and supplier concentration increasing. Examined separately, each is explicable on its own grounds. Examined together, they are the signature of a system in which units optimize their own measures against one another's cost base. The working capital effect typically becomes visible one budget cycle late, since an inventory build can be reported in the first quarter as an improvement in service levels and, presented that way, reads as a favorable development rather than a warning.
In capital-intensive projects the same disconnect is priced through the schedule rather than the balance sheet. Where the critical path in the works programme does not coincide with the metric set held at the executing layer, the gap between progress reporting and the actual commissioning date compounds; at the point where that gap touches the liquidated damages cap, the completion date in the credit agreement, or the supply start committed to the offtaker, the cost of correction ceases to be an engineering cost and becomes a contractual one. When renegotiation opens, the bargaining position of each party depends on whether the origin of the delay can be documented as falling within a particular party's control, and in an organization that kept no deployment record, that documentation does not exist.
At the valuation desk the effect is more indirect but considerably more durable. Where a diligence process cannot establish a traceable link between the company's metric set and its stated strategic priorities, the reasonable inference available to the buy side is that results originate in the judgment of specific individuals rather than in a system; that inference is typically recorded under founder dependency and expresses itself not as a headline price discount but as structure — an earn-out, key-person undertakings, or an expanded representations and warranties package with a correspondingly larger escrow. What determines a multiple is, more often than not, not past performance itself but the demonstrability of that performance being repeatable independently of the founder, and the deployment record is the cheapest available instrument of that demonstration.
The mechanism that neutralizes this tendency is not individual discipline but an institutional architecture with four separable components. The first is binding the number of priorities to a ceiling consistent with actual resource allocation, and writing down explicitly which priority is displaced whenever that ceiling is breached. The second is recording, at each layer, which metric the objective was translated into and on what assumption that translation rests; the critical artifact is the assumption rather than the metric, because it is the assumption that lapses. The third is mapping horizontal dependencies — which unit's target depends on which unit's output — at the moment targets are set, not at the moment they are missed. The fourth is structuring the review rhythm to interrogate the cause of deviation rather than the metric result; why a measure is green carries as much information as why another is red.
BEIREK's intervention in capital-intensive projects and portfolio transformations is built on exactly these four components. Starting from the return and schedule commitments made at sponsor level, it constructs an objective-to-metric mapping matrix that descends through the project company, engineering management, procurement and site layers, holds the capacity and lead-time assumption underlying each mapping in a separate column, and retests the validity of those assumptions not on a calendar quarter but against the project's own milestone rhythm — signing, closing, first draw, commissioning. Before objectives are pushed down, the receiving layer's feasibility objection and the manner in which that objection was addressed are entered into the record, so that a deviation emerging mid-year opens into a question about which assumption failed rather than a dispute about who is accountable.
The second line of intervention is the contractualization of horizontal interfaces. Dependencies between units and between contractors, rather than being left to goodwill and a weekly coordination call, are written at the level of who owes which output to whom by which date; once the layer at which a delay originated becomes traceable through that record, correction engages earlier and the documentary position a party carries into a liquidated damages or variation negotiation forms of its own accord. The secondary and generally underestimated return on this record set is that, in a sale or financing process, the diligence team's question — whether the result comes from the system — can be answered with documents rather than narrative.
Whether a strategy was executed is usually established not from the year-end results pack but from whether the assumptions under which that strategy was translated, layer by layer, were ever written down; absent that record, every metric delivered proves only itself.