Owning the Conversion from Strategy to Work

Nobody refuses the strategy. Each function receives it, reads it seriously, and converts it into the part it can act on with the resources it controls. Five competent conversions, no two requiring the same thing to happen.

That is the failure, and it does not look like one. It looks like a year of accurate green status reports and a result that never arrives, with no line in any report marking where it was lost.

What actually happened

Polymath’s founder held the Office of the CEO in that group for three years: first as vice president, then as senior vice president and Chief of Staff to the chief executive with global data and insights attached. A consumer packaged goods business, roughly $2–3 billion in revenue, listed on the TSX after years in private hands. The work was board operations, board-level business health reporting, and enterprise priorities running across finance, HR, technology and commercial. Strategy and the numbers underneath it sat on one desk. That is the only reason what follows was visible.

Coming out of the pandemic the company carried too much stock in the wrong shape, and inventory optimization was one of the priorities he ran. One sentence in the plan read as an instruction to hold less inventory without losing service. Everyone in the room had agreed it.

Follow it into the operating plans and it stops being one sentence. Finance held it as working capital released by year end. Supply chain held it as cover against stockouts, which means more of the right thing rather than less of everything. Commercial held it as never being short for a major account’s shelf reset. Sourcing held it as unit cost protected, and unit cost is bought with longer runs. Product held it as range depth, because breadth sells a category in. Five plans, one cycle, each correct from where it was read.

None of them was the commitment. The commitment was the trade-off between them: how much unit cost the company would pay to release cash, how much service risk it would carry, which accounts were protected and which were not. No plan contained that, because no function could make it. Sourcing cannot concede unit cost for the margin line.

What the Office of the CEO did was not sophisticated. It put each function’s version of the sentence on one page, side by side, and took it to the executive team as a decision rather than a status update. The chief executive settled three incompatibilities in the room and refused the fourth until sourcing and finance returned with the cost of both. The conversion itself was not difficult. It had simply never been anybody’s job, so it was done five times, privately, by people who could each see only their own half.

Polymath puts a client’s five private conversions on one page and takes them up as a decision, and the harder half is holding one named person to that job afterwards.

Why it happens

Strategy is written above the functions, because that is the only altitude at which its sentences are true. Hold less inventory without losing service is coherent for an enterprise and incoherent inside any function, where it becomes two objectives and one belongs to somebody else.

Execution only happens inside functions. A function is the smallest unit with a budget, a calendar and a person who can be held to a result. The enterprise has none of those, so you cannot assign work to it.

Between the two sits a specific piece of work: deciding what a commitment requires of each function, which existing objectives it displaces, in what sequence, and what each concedes so another can deliver. It takes weeks. It also has no owner, structurally rather than carelessly. Budgets follow functions, and no leader can be held accountable for what a peer gives up.

So the work falls to the only people with authority to act. Each function converts the commitment itself, in the frame it is measured in, keeping the portion it controls. What drops out is the portion requiring somebody else to move — which is why the commitment was set at enterprise level.

A strategy divided among functions has not been delegated. It has been reduced, five times over, and every reduction is defensible.

Nothing signals this while it happens. Each function is delivering its own conversion, so each reports green, and each is telling the truth. There is no variance to point at: the loss occurred at conversion, in five meetings that never met. When two conversions do collide, they collide late and in the worst venue: the executive meeting, as a delivery problem between two credible leaders, each holding evidence, each correct inside their own frame. The person who can adjudicate has the least detail and twenty minutes. The decision is split, or deferred, which is the same decision.

Then the plan is re-based. By the second re-basing nobody can separate commitments that were tried and proved wrong from those never converted into anything to try. The strategy stops being tested and starts being replaced.

The strongest objection

The strongest case against this: we already have a PMO, and this is precisely what it does.

A competent PMO does the visible part well — dependencies mapped, dates challenged, slippage surfaced while it is still cheap. What it lacks is standing. Its authority derives from a plan produced by the conversion in question. It can record that two functions disagree; it cannot decide between them. When a function’s reality contradicts the plan, its only instrument is a re-baseline, which records the loss.

The second objection is harder: this puts a person between you and your executive team, and if your leaders cannot convert a strategy into their own function’s work, you have the wrong leaders.

The first half is a genuine risk. An Office of the CEO that accumulates informal authority becomes a filter and holds decisions it was never granted; the design we propose below exists to prevent that. The second half misreads the failure. Each of those five conversions would survive scrutiny in its own functional review. Standing was the constraint, not capability: no function head can bind a peer’s plan, and none can see the other four conversions at the moment theirs is made. The role owns no function and no result, and everything it produces is visible.

The method

We run five steps, in this order, and we do not drop the middle two: drop them and you have a coordinator.

1. We write every enterprise commitment as a sentence naming what is given up. We do not accept one that contains no concession: it has not been converted, it has been announced. We turn hold less inventory without losing service into a sentence stating the unit cost you will accept to release cash, and which accounts are protected ahead of which. The trade-off we accept: the concession sits where everyone can see who pays it. Skip it and each function converts the half that costs it nothing.
2. We assign the conversion to a named person, and we call the role what it is: Chief of Staff. Not the diary and inbox version of the title, which is why we specify it in writing. A full-time role in your Office of the CEO whose product we define as the converted commitment set: what each function must do, stop, sequence and concede. We have it report to the chief executive, not the COO and not the PMO, because it has to tell a function head what no peer can. The trade-off we accept: the executive team reads it as a layer before it reads it as a service. Skip it and the work falls back to the functions that cannot do it.
3. We give the role decision rights in writing, with an explicit ceiling, and we write the ceiling as carefully as the rights. It may require each function to publish its conversion in a common format by a date, return one as incomplete, settle conflicts that change sequence, interface or priority, and frame an item for the executive agenda, deferrable once. It may not change targets or budgets, direct people, or own delivery. The trade-off we accept: your chief executive must back the role in public the first time a right is used against a senior leader. Skip it and you have a coordinator with nothing but the CEO’s calendar.
4. We publish the conversions side by side, with what each function will stop doing. We produce one page per commitment, one column per function, circulated before delivery begins, each column naming the work displaced. We let the incompatibilities become the agenda instead of a discovery, while changing an answer still costs a revised plan rather than a written-off quarter. The trade-off we accept: the first session is uncomfortable, because an executive team is disagreeing about something they all signed. We sit in it and chair it anyway. Skip it and you meet that disagreement at quarter end, attached to a variance.
5. We time-bound the role to the commitment set and re-grant it every year. We let the role expire with the commitment set it converts, renewed deliberately or not at all. The trade-off we accept: you lose accumulated context at every reset, and we would rather lose that than let the role harden. Skip it and the Office of the CEO acquires staff, a process and a standing claim on decisions belonging to your executives — the failure the second objection predicts.

Notice what we have not added. No committee, no cadence, no system. The conversion is a job, and this is what we have found it takes to hold one person to it. It is also the lifecycle compressed into a single role: we diagnose where the conversions diverge, architect the commitment set, redesign the routines carrying it, implement, enable your functions to run it unaided, and optimize. Each step runs sideways through a company built, budgeted and promoted vertically. Organizations optimize what they own. The value in this commitment existed only between Procure-to-Pay and Order-to-Cash, and nobody owns the space between two value streams.

Evidence and measures

One commitment — hold less inventory without losing service — as six teams read it:

Five measures worth tracking over ninety days. None requires a new system:

— For your top five commitments, ask each function for its written interpretation. Count how many arrive in a week, then how many agree.

— How many of those commitments have a named person accountable for the conversion rather than the outcome. Expect zero; that is the finding.

— Elapsed time from a function discovering its plan conflicts with another’s to that conflict reaching someone who can settle it.

— Share of cross-functional items reaching the executive meeting framed as a decision with options rather than a status problem.

— Executive decisions deferred twice. Twice is the signal; once is a reasonable request.

If those five move, the conversion has an owner. If the only change is more alignment sessions, it does not.

How Polymath solves this

We start with a conversion audit, and it takes about a fortnight. We ask each function for its own written interpretation of the same handful of enterprise commitments, in its own units, on one page, by a date, and we lay the versions side by side. The divergence is the diagnosis, and it is wider than your executive team expects, because nobody has seen a version but their own. This is the one piece of work we run against all four of our disciplines rather than one, because the conversion touches how the enterprise is architected, how its processes run, how its systems record the answer and how its people are held to it. That is why we sell it as a single engagement rather than four, and why we will not split it into four.

We then rewrite the commitment set with the concessions named, in working sessions with your executive team rather than in a report to it, run by people who have held this from inside an Office of the CEO. We build it into what you already operate: the review agenda, the board pack, the decision log, the format your functions file in. We chair the first cycles jointly, and then your Chief of Staff chairs them and we step out.

What compounds is the routine. Once a commitment set exists in converted form, next year’s strategy meets a structure that already knows how to take it apart — the format, the escalation ceiling, the decision log, the definitions your reporting runs on — so your people convert the next set in weeks rather than a planning cycle, and the horizontal work stops depending on who stands nearest the chief executive. What a later executive team inherits is a conversion already argued through once, so the argument is about the trade-off itself rather than about who owns making it.

What it costs to do nothing

Strategies that fail loudly are cheap. They get cancelled and the capital comes back. The expensive one is never rejected and never delivered: it consumes a full planning cycle of the scarcest attention in the company and returns five defensible plans that partly cancel out.

The trade-off still gets settled, incidentally, by whichever function has the strongest reporting line — usually the one closest to the period-end number. Cash is released and the unit cost surfaces two quarters later on a margin line with no path back to the decision. Or cost is protected, the cash release never happens, and the explanation offered is demand.

The compounding cost is credibility. A board watching the same commitment appear in a third consecutive annual plan has stopped evaluating the strategy. It is evaluating whether this team can execute one, and in a public company that judgment gets priced.

Problems are solved vertically. Value is realized horizontally. The horizontal work is a job, and until a name sits against it, it stays everybody’s second priority. Putting the five conversions side by side is where we begin.