One Surface to Argue From

A strategy split across three decks is three strategies.

If you've watched a strategy get approved in three separate documents, and then watched two functions discover months later that they'd agreed to different things, nobody in that story was careless. The contradiction was there at approval. It was simply not in a form any human could have seen, and there's a measured limit that explains why.

Strategy is the subtraction, and three decks are how you avoid it

Michael Porter defined strategy in Harvard Business Review in 1996 as the choice of what not to do. Not a list of everything worth doing, which is an inventory of ambitions, but the deliberate decision to give some good things up so the rest can concentrate.

The essay is quoted constantly and practiced rarely, because subtraction is uncomfortable in a way addition never is. Saying no to a good idea, in front of the person who proposed it, is harder than saying yes to everything and sorting it out later.

Three decks are the mechanism by which organizations avoid the subtraction. Nothing has to be surrendered, because each document keeps its own ambitions and never has to meet the others. No one lies and no one dodges. The format simply never forces the conversation.

Why nobody catches it, even reading carefully

This next part takes the blame off the reader.

Working memory holds roughly four items at a time. That's Nelson Cowan's estimate, defended in Behavioral and Brain Sciences in 2001, revising George Miller's famous "seven, plus or minus two" downward. Four.

Now consider an operating model with fifteen roles, eleven lifecycle phases, and three account tiers. Reconciling that across three separate documents requires holding far more than four things at once, so nobody does it. People reason within whichever deck is open and trust that the others agree. That isn't a discipline failure. It's the predictable consequence of asking a person to do something the architecture of human attention doesn't support.

Which means the fix isn't better reviewers or more careful reading. It's a format in which the contradiction is visible in a single view.

What one surface actually did

The operating model I was asked to design for a global facilities-services company's new account-management program could have shipped as three decks: one for roles, one for the service lifecycle, one for account tiers.

It shipped as a single surface. Fifteen or more roles down one axis, eleven lifecycle phases across another, three account tiers layered through it, and over the top an overlay tracking how each team was meant to feel at each phase. How the people running a program experience it is an operating variable, not a soft one, and a program fails wherever that variable is left off the model.

The single surface wasn't a formatting preference. It was the intervention. Three decks would have let each function keep its own version of the program and reconcile the differences later, or in the meeting where they collided in front of a client. One surface made the contradictions visible while they were still cheap to fix: a role on the first axis with no home in any phase on the second, a tier promising a level of service the roles couldn't staff.

The artifact didn't describe the strategy so much as decide it, because two incompatible ideas could no longer sit on separate slides and appear to agree.

The deliverable is the intervention

This is the part that travels past one program.

Most deliverables describe a decision that was supposed to happen somewhere else, in a workshop or a steering committee or an offsite. The good ones are the decision. When an artifact is built so its own structure forces the choices, completing it and making the strategy are the same act, and there's no later meeting where the real call gets made, because the real call was made in the room where the surface was drawn.

A deliverable that only records a strategy can be immaculate and change nothing. A deliverable that constitutes the strategy can't be finished until the organization has actually chosen. The first kind is safer to produce and easier to approve, which is exactly why most of what gets commissioned is the first kind.

Scope for what it forces

There's a question worth asking before you commission anything, and it takes one sentence.

What will this deliverable force?

If it can be completed without anyone giving anything up, it will be, and you'll have bought a description of a strategy rather than a strategy. If you want the smaller version of this, take the two axes of your current plan that are most likely to conflict and put them on one page, side by side, before the next review. That's usually enough to surface the argument you've been having in fragments.

Build the one surface that can't be finished until the contradictions are resolved, and hand people the version of the work that does the deciding rather than the version that waits for someone else to do it later.

---