The Analytics Group

Illustrative. The people and the companies are invented.

Cresterly Industrial is a publicly traded $4 billion manufacturer of high-precision motion-control equipment. It employs about 6,200 people, runs eleven plants across three countries, and has been growing slowly but consistently for a decade. Its executive committee comprises a CEO (Patricia Lin, two years in role, hired to accelerate growth), a CFO, a COO, a Chief Operations and Manufacturing officer, a Chief Commercial Officer, a Chief Technology Officer (Marcus Vale, four years in role), and an SVP of Operations Planning (Diana Reyes, who reports through the COO).

Reyes runs operations planning, demand forecasting, supply-chain optimization, and inventory strategy. Her group has been incrementally adopting analytical tools, but she does not control the company's central data infrastructure or the firm's 40-person Data & Analytics group. That group, which builds the data platform, runs the predictive models, and partners with business units on dashboards and reporting, sits under Vale in Technology.

Reyes has come to believe that the analytics group is structurally misplaced. In her view, the closest customer of advanced analytics is operations; the work would be faster, better targeted, and less politicized if it reported into her function. Vale, predictably, disagrees: data unification is a CTO responsibility; fragmenting it under business-unit leaders would balkanize the platform and create governance problems. Lin, the CEO, has thus far been agnostic but is publicly committed to "data-driven operating discipline" as a margin initiative.

This is the setting in which Reyes will run three cycles of The ADAPT Loop.

The probe

Reyes's actor profile is clear. Her agenda: become the company's center of operational gravity, on a trajectory toward COO and then CEO. Her interest this quarter: establish a credible claim that operations is the home of analytical decision-making. Her objective: secure a defensible operating relationship with the analytics group, short of full absorption. Her goal: a chartered joint pilot program in which her function and the analytics group co-deliver a forecasting improvement to a single business unit. Her target: Lin's eventual decision to recognize operations as the primary internal customer of analytics work.

Her resource check returns Yes. She has political capital with Lin from a successful inventory-reduction initiative last fiscal year. She has the CFO's quiet support; the CFO does not trust Vale to drive cross-functional financial outcomes. She has sufficient time and attention to run a pilot herself. She has a plausible business case: a 6 to 8 percent margin improvement in the targeted business unit, conservatively estimated.

Her options are not infinite. She considers three, and she scores each option.

The matrix tells her something useful. Option A has the highest preferred-position utility, but its adverse alternative is catastrophic: a frontal refusal would set her back twelve months and harden Vale's position. Option C has acceptable downside but no upside motion. Option B is the workhorse: a modest preferred-position payoff, an essentially neutral adverse-alternative outcome. She selects B.

She acts. The action's components: a one-page proposal co-authored with the business-unit GM whose unit the pilot would target; an informal hallway conversation with Lin three days before the proposal lands; a deliberate copy-in to Vale to forestall accusations of going around him.

The actual effect lands closer to Reyes's adverse alternative than her preferred position. Lin approves the pilot, but with a steering committee chaired by Vale "to ensure data integrity standards." Reyes is in the room, but the gavel is not hers. Her actual utility is roughly +1 (lower than her expected +4). Her perceived utility, however, is closer to +3: she overreads her presence in the room and underweights the chair assignment.

The perception gap is +2. This is the error term she will carry into the second cycle without realizing it.

The resource loop drains her slightly: she has spent some political capital with Lin to get the pilot approved. The learning loop updates her interests subtly. She now sees Vale's procedural skill more clearly, and her interest shifts from "claim the analytics group" toward "establish a track record that makes the claim self-evident."

The disciplined hold

Reyes returns to the funnel. Her agenda is unchanged. Her interest has updated: she now wants to demonstrate, over the course of two quarters, that operations can drive analytics work in a way the existing structure cannot. Her objective tightens: deliver, through the pilot, a visible operating result that Lin and the CFO can point to as a margin proof point. Her goal becomes the pilot's quantitative outcome, due in the spring. Her target shifts from "Lin's decision" to "Lin's narrative about the firm's operating model."

The capability gate now returns No. She has spent political capital, the pilot is consuming attention bandwidth, and the next move (pushing for a formal restructuring proposal) would exceed her current resource balance. The framework's discipline pays off: she does not push. She rescopes. The new objective is to maximize the pilot's quantitative result and to build a quiet coalition with the CFO and the targeted business unit's GM, so that any future restructuring discussion is initiated by someone other than Reyes.

This is the part of the framework that most practitioners skip. The No-loop is not a failure state; it is the most strategically sophisticated position the framework can produce. A disciplined hold, with deliberate coalition-building underneath, is what permits the third cycle to be decisive rather than premature.

While Reyes holds, something happens on Vale's side that she does not yet see, and that the framework will only fully explain in retrospect. Vale begins a major cloud-infrastructure migration in his Technology organization, a multi-quarter, high-cost program that consumes most of his attention. His own observation node, which had been alert to Reyes in the first cycle, becomes starved. He is no longer watching the board on which Reyes is moving.

The strike

Six months later, at the company's spring offsite, Reyes runs the cycle for the third time. Her resources have rebuilt. The pilot delivered a 7 percent margin improvement in the target business unit, on the upper end of her conservative estimate. The CFO has begun publicly attributing the result to "an operating discipline that ran from the demand model through to the work order," language Reyes herself drafted. The business-unit GM has asked, in two executive meetings, whether the model could be extended to his peer units. Vale has missed two committee meetings due to migration overruns. The CFO, in a quarterly business review, has questioned the migration's slipping timeline.

Her options have changed. The frontal option that was reckless in the first cycle is now viable.

She scores it: U(PP) = +7, U(AA) = −1. The adverse alternative is acceptable because the pilot's results will not disappear; the case for restructuring will be available again in the next planning cycle. She selects D.

She acts. The components: a memo authored with the CFO; a pre-aligned conversation with the business-unit GM, who has agreed to second the proposal; a deliberately collegial framing toward Vale ("data platform integrity stays with Technology; only the applied layer moves").

The actual effect lands close to her preferred position. Lin approves the split, with two conditions: a dotted-line relationship to Technology to preserve platform integrity, and a 12-month review at which the arrangement will be reassessed.

The reader who has been paying attention will notice that the leash (the dotted line and the review) is the visible trace of Vale's matrix in the third cycle. He had time to run only one cycle, late, under pressure, with a depleted resource base; his preferred position (block the split) was unaffordable, so he selected the adverse-alternative move (concede the applied layer; entrench on the platform; extract a review). Both actors achieved their selected options. The outcome that looked, from Reyes's perspective, like the CEO hedging her bets, was in fact the resolution of two actors' decision matrices through a shared target.

Lessons

The matrix's discipline is in the adverse alternative, not the preferred position. Reyes's first-cycle selection of Option B over Option A was driven by AA, not PP. A practitioner who scores only preferred positions will systematically overbet.

The No-loop is a strategic asset. The second cycle is the cycle that most practitioners skip and most situations require. Holding deliberately, with continued coalition-building underneath, is what allowed the third cycle to be decisive.

Apparent randomness in outcomes is usually another actor's matrix. What looked, from a single-actor perspective, like uncontrollable reality (joint governance, the leash, the review) was, in the two-actor view, the precise output of Vale's matrix. There is rarely as much noise as it feels like there is.