The document on your desk is good, and you have said so out loud. It surveyed a market you thought you knew, found three things you did not, and named the decision: two segments to leave, one to fund properly, a price change you had been circling for a year. The reasoning holds up under argument, which is why you agreed with it in the room. That was Thursday. It is now Tuesday, the deck is a link in an email, and your commercial director is answering messages about last week. Nothing has gone wrong. The decision has arrived somewhere with no free hands.
Deciding and doing are staffed differently
Begin with what a deck does that nothing else on this page can do. It decides what should be true. Good advisers survey what you cannot survey from inside, reason in the open, and put a recommendation in front of you that adults can argue with in a room. Software cannot do that. An agent holds no view on which segment to leave, cannot interview your customers about a competitor they mentioned once, and has no standing to tell you your pricing is wrong. When a company does not know what it should be doing, a document that settles it is the most valuable purchase of that quarter.
What the document does not have is hands. The gap it lands in is a distance of capacity: between a decision made and the standing obligation to carry it out through weeks nobody planned.
The Cabinet, ArkOne’s reference design for an executive agent, is built for that distance rather than for the decision. Each specialist carries goals in one of three states, on track, at risk or off track, reported on a cadence without being asked. The Clock wakes every thirty seconds and does what it was told whether or not anybody remembered. The Minutes are decisions written after the work and read back at the next start, the eight most recent alongside every open initiative. None of that is thinking. It is the machinery of not forgetting, which the deck assumed somebody would supply.
The six rows
Suppose a specialist manufacturer of four hundred people, invented for this exercise. A capable advisory firm spends ten weeks on its route to market and delivers a recommendation the board endorses without dissent. Six months later the price change is live and the two exits have not begun, because the exits need a hundred conversations with distributors and the person who would have had them is running the business.
| The row | A strategy deck | The Cabinet |
|---|---|---|
| Who decides | Your board, once, on the evidence the document assembles; then nobody again | A specialist at its authority setting, on each case as it arrives: act, propose and wait, or escalate. All ship at propose |
| What it can touch | Nothing. It persuades people, who then touch things | Named systems through named tools, approval rights written per person, and a plan streamed before the work so you can stop it |
| What it remembers | Whatever the reader remembers; the file never changes and never consults itself | Decisions written after the work and read back at the next start, plus open initiatives and a morning brief that synthesises rather than re-lists |
| What it costs to run | Nothing after delivery, so no bill ever arrives to show that nothing is happening | A running bill that follows the work, and rises only while the work continues |
| What happens when it is wrong | The recommendation is quietly not pursued, and the reason is rarely written anywhere | An append-only entry naming the action, the reason and who approved. A correction is a new entry and the original stands |
| Who owns it afterwards | The sponsor, until they change role; then the file has an author and no owner | Instructions a policy owner edits as text, the authority settings, the roster and the ledger, each with a named owner or none |
Exhibit 1. Illustrative. Six rows separating the document that decides from the thing that carries the decision.
The fourth row is the one nobody plans for. A delivered document costs nothing to keep, so a strategy going nowhere is silent: no invoice, no failing job, no line queried at review. A machine that bills while it works is an irritation and also a signal, because the bill stops when the work does, and somebody notices that in a month rather than a year.
The third row is where the six months go. The document is fixed and everything about implementation is not: a distributor changes hands, a competitor moves, the person leading the exit is on parental leave from March. The deck cannot re-read itself in that light. A person can, and that person has a job, so the re-reading is what gets skipped, and it is the whole of implementation.
The first row deserves an honest reading in both columns. Concentrated deciding is a strength of the deck rather than a weakness: a board should settle a direction once, deliberately, with the evidence in front of it. What follows is hundreds of small decisions nobody convened a meeting for, and those are the ones that go unmade.
When the thinking is the thing to buy
Two situations make advice plainly the right purchase, and the first is the more common.
A company that cannot say in one sentence what it is trying to change should buy the thinking and automate nothing yet. Machinery built around an unclear intention produces activity, reporting and cost, and no progress, and unwinding it is harder than never building it. If your executive team would give three different answers to what the next year is for, that disagreement is the work, and it is resolved by argument in a room rather than by any system.
The second is a genuinely novel question. Entering a market nobody in your company has operated in, pricing something with no comparable, judging whether a regulatory change is a threat or an opening. These need somebody who has seen the shape before, and no ledger or cadence substitutes for it.
The question that separates the situations is about your last three quarters rather than about your ambitions: is there a decision you have already made, agreed with, and failed to carry out more than twice?
A single failure is circumstance. Three is a capacity finding, and it says the deciding was never the constrained part. More analysis then buys a better version of what you already hold, and it lands on the same Tuesday.
Take the last decision your board endorsed and find who holds it in writing as their work this week, with a date. If somebody does, you have a delivery problem you can manage. If everybody merely agrees with it, the decision has no owner and the document cannot become one. The Cabinet and an outsourced team puts the same ownership question to a supplier, and every setting cited here sits on the Register.
component: null
Asked plainly
Why do good strategy recommendations so often go unimplemented?
Because a recommendation is a decision and implementation is a standing obligation, and the two are staffed differently. The document names what should be true and the people who must make it true already hold full jobs. Nothing in the document wakes anybody up, chases the item that slipped, or reports that a target moved. Where implementation succeeds, somebody was given the work as their work rather than as an addition to it.
Do we need consulting advice or software?
It depends on whether the open question is what to do or whether it is getting done. A company that cannot say clearly what it is trying to change should buy the thinking, because automating an unclear intention produces motion and no progress. A company holding a decision it agrees with and has failed to carry out three quarters running has a capacity problem, and more analysis will not touch it.
What actually carries out a strategy after the advisers finish?
Usually a named owner, a cadence somebody protects, and a written record of what was decided and what has moved since. Where those three exist, implementation tends to happen with ordinary tools. Where they do not, the work depends on whoever remembers it in a busy week. Software can hold a cadence and a record reliably, but it cannot supply the owner, and a decision with no owner does not survive contact with a quarter.