Good Planning Bad Planning
The Lost Discipline between Choice and Action
1. A strategist’s confession
I have spent years inside strategy.
The Strategy Choice Cascade. Activity systems. Where to play, how to win, what capabilities to build, what management systems to install. I have read Porter, Martin, Rumelt, Helmer, Christensen, Freedman, Carroll & Sorensen, Ghemawat, Rosenzweig… The canon, cover to cover, and I deeply believe that strategy is the most critical discipline in management. Getting strategy right is hard, rare, and consequential. A genuinely distinctive competitive position, backed by integrated choices that produce real trade-offs, is among the most valuable things a leadership team can create. I have built a career helping organizations make those choices.
And here is what I have learned: the problem of strategy is even harder than it looks.
Most organizations cannot do strategy at all. What they call “strategy” is a set of financial targets, or a budgeted initiative list, or a vision statement, or a competitive benchmarking exercise. Anything, except an integrated set of choices about where to play and how to win that would cause a competitor real pain to replicate. I have written at length about this confusion, about the parasitic meanings that have colonized the word “strategic” and drained it of content. Strategy remains, in my view, the most pressing problem in management. Getting it right is rare, and most organizations fail here.
Roger L. Martin has been saying this for years, and saying it better than anyone. His article “Strategy vs. Planning: Complements not Substitutes“ remains the cleanest boundary-drawing in the business literature: strategy is an integrated set of choices that positions the organization to win; planning is projects with timelines, deliverables, budgets, and responsibilities. In “Why Planning Over Strategy?“, he diagnoses the disease with clinical precision: a self-reinforcing system in which business schools produce technocrats, technocrats populate consulting firms and corporate strategy departments, and all of them prefer analysis-disguised-as-strategy to the harder work of actual choice. Richard Rumelt is even more direct. In “Becoming a Strategist, Part 1“ he writes: “ignore ‘strategic planning.’ It is almost always financial forecasting, not strategy.”
I agree with all of this. The diagnosis is correct. Strategy, as a discipline, is dying under the weight of a planning process that has annexed its vocabulary and displaced its intellectual substance.
But a question began to gnaw at me.
Assume you solved the strategy problem. Assume you were one of the rare leadership teams that actually made genuine strategic choices: a distinctive where-to-play, a defensible how-to-win, real trade-offs, a coherent activity system, the kind of integrated choice Martin and Rumelt spend their careers defending. What would that strategy find downstream? What organizational machinery would receive it and convert it into coordinated action?
The answer I kept encountering, in organization after organization, sent me on an investigation that I did not expect to take me where it did. Because the gap between a brilliant strategic choice and the organizational machinery that was supposed to make it real turned out to be a missing discipline with its own canon, its own architecture, and its own deep thinkers. A discipline that has no bookshelf.
Nobody wrote Good Planning Bad Planning.
2. Where Martin and Rumelt stop
Before going further, I want to sit honestly with what Martin and Rumelt actually say about planning, because it matters for where this essay goes.
Martin has written at least five substantial pieces on planning over five years. In “Strategy vs. Planning“, he draws the boundary and insists the two are complements. In “From Strategy to Planning“, he describes planning as communicating your choices, chartering subordinates’ choices, and following up: “turtles all the way down.” In “Why Planning Over Strategy?“ he attacks planning as a colonizer of strategy’s vocabulary. In “Integrating Strategy with Planning & Budgeting Processes“, he makes the crucial point that budgeting enjoys “possession” and that strategy must be proactively integrated into the budgeting process or it will not happen. And in his recent “All-Stars Book Club Chapter Two“ reflection, he doubles down on the technocrat-versus-strategist frame. This is an enormous body of work.
Now notice what Martin does and does not do with it.
He draws the boundary between strategy and planning with unmatched precision. He identifies the political and educational forces that make planning dominant. He describes, correctly, how budgeting captures spending authority and therefore must be force-integrated with strategic choice. He gives working definitions: planning is “projects with timelines, deliverables, budgets, and responsibilities.”
But he does not, in any of these pieces, write about planning itself as a discipline with its own internal architecture. He treats planning primarily as a thing that must be managed in relation to strategy rather than as a field with its own thinkers, schools, failure modes, and craft standards. Something to be subordinated, integrated, disciplined.
But then, what distinguishes good planning from bad planning, once you grant that it is complementary to strategy rather than substitutive? What does the internal structure of a serious plan look like? What does one great planner know that another does not? Martin’s body of work does not answer those questions, because those are not the questions Martin is working on. He is working on strategy, and planning appears in his writing mostly as its necessary, sometimes threatening neighbor.
Rumelt goes further in the dismissive direction. “Strategy is not planning. It is not simply setting goals,” he writes in “The Shade of Strategy.” The “simply” is telling. Planning, in Rumelt’s writing, is the thing strategy is emphatically not. Not something worthy of its own canonical treatment.
I agree with both of them on the boundary. I agree with both of them on the urgency of defending strategy from technocratic colonization. I admire Martin deeply; his work on strategic choice has shaped how I practice. But there is an unfilled intellectual space, here. If real strategy is rare, and if rare strategy still has to pass through the organizational machinery that converts choice into action, then the craft of that conversion should deserve the same depth of treatment strategy has received. Nobody has given that treatment to the discipline of planning itself, done well.
And once I noticed the gap, I could not stop noticing what sits inside it. If Strategy is a Lost Art, then Planning is a Lost Discipline.
3. The planning I keep seeing
Here is what sits inside the gap.
The leadership offsite has ended. A strategy has been agreed; let us imagine a genuinely good one, the kind Martin would recognize as an actual strategic choice. A distinctive where-to-play, a coherent how-to-win, real trade-offs made at real cost to other possibilities. Everyone shakes hands. Someone assigns the head of strategy, or the CFO, or a newly anointed transformation office to “build the plan.”
Six weeks later, the plan arrives.
It is a slide deck, usually between thirty and seventy pages, with an executive summary on page two. The summary contains four or five strategic pillars, re-stating the strategy one-pager at “a more concise manner,” phrased at exactly the level of abstraction that feels meaningful in a boardroom and means nothing at 9 AM on a Tuesday when someone has to decide whether to hire a second product manager. Under each pillar sits a list of initiatives. Fifteen, twenty-five, sometimes forty of them across the whole plan. Each initiative has an owner, usually a function head who did not know they were being assigned until the plan was distributed. Each has a timeline, rendered as a Gantt bar across three years. Each has a preliminary budget envelope, negotiated in the last two weeks before the deck was finalized by backsolving from the finance team’s aggregate target, to be re-confirmed during actual budgeting.
The financials are at the back. This is the part everyone flips to first.
The financials are presented as a three-column table. The columns are labeled, in bold, in exactly the way they have been labeled in every planning deck in every industry in every Western economy for the last thirty years.
Base case. Optimistic case. Pessimistic case.
In the base case, revenue grows eight percent a year, margin expands by thirty basis points, and the company hits its strategic milestones on the schedule shown above. In the optimistic case, revenue grows eleven percent, margin expands by fifty basis points, and the milestones are hit slightly ahead of schedule. In the pessimistic case, revenue grows five percent, margin compresses by twenty basis points, and the milestones slip by a quarter or two.
The three cases run neatly across the page. They produce three different EBITDA numbers. They produce three different cash-flow profiles. They are accompanied by a narrative paragraph each, explaining what “macro tailwinds” or “execution challenges” might drive each scenario.
Now look at what is beneath the financial table: the initiative list, the resource allocation, the capability investments, the hiring plan, the capital expenditure calendar.
They are identical across all three scenarios.
The same initiatives run. The same resources are allocated. The same people are hired on the same timeline. The same capital is spent. In the “optimistic” case, the organization does exactly the same things it does in the “base” case. In the “pessimistic” case, it does exactly the same things. The three scenarios are three different numerical outputs produced by the same decision set. They are not three different scenarios. They are one scenario dressed in three costumes.
Nobody notices this. Or rather, everyone notices it and nobody names it, because naming it would require someone to stand up in the room and say: what, then, is the point of the scenarios?
I have sat through this meeting more times than I can count. I have watched smart, well-paid executives nod gravely at the pessimistic case, ask one or two questions about the optimistic case, approve the plan, and leave the room satisfied that the organization has exercised due diligence on uncertainty. The scenarios performed the function of making the plan look thoughtful. They changed nothing about what the organization would actually do.
And that is not even the worst of it.
Go back to the initiative list itself. Look at how each initiative is described. Most of them are stated as commitments: launch the new platform by Q3, open three distribution centers in the Southeast. But a significant number are actually hypotheses about things the organization does not control: customer conversion will rise from twelve percent to seventeen percent; churn will remain below three percent; the regulator will approve within ninety days. These hypotheses sit on the same Gantt chart as the commitments, with the same green status indicators, reviewed in the same monthly cadence as if they were completable tasks.
When a commitment slips, it is a project management issue. When a hypothesis fails, it is usually treated as bad luck, an “unforeseen market condition,” rather than what it actually was: a bet the plan had no machinery to monitor or adjust.
Now ask yourself: what are the two or three load-bearing assumptions that, if wrong, would invalidate the entire plan? In the plans I have seen, those assumptions are either absent from the document entirely or buried in the footnotes of page forty-three. There is no signpost system. No trigger point. No pre-authorized reallocation protocol. No contingency branches for upside or downside. If demand surges past the base case by thirty percent, there is no pre-designed response; the organization will scramble. If demand collapses by twenty percent, there is also no pre-designed response; the organization will also scramble. The plan has one mode: execute-as-written until reality breaks it, then meet in a conference room.
This is what I mean by “bad planning.” Not that the planners were lazy or stupid. Most of them are extremely smart and working extremely hard. The problem is that the document, the cadence, the scenarios, the review process they are producing are structurally incomplete as a planning system. It has project schedules. It has a budget. It has something called “scenarios.” What it does not have is a mechanism for making the strategy survive contact with money, capacity, time, and surprise.
And every single organization I have seen produces roughly this same document. The industry does not matter. The assisting firm does not matter. The CEO’s pedigree does not matter. The pattern is so consistent that it must be produced by something more fundamental than local incompetence. It is produced by a discipline that was never articulated, so nobody knows what it should contain. When you have no canon, everyone copies the nearest available template. And the nearest available template is the one that every other planning presentation uses, because that is the one every planner was trained on.
That was the observation that started my investigation.
4. Seven disciplines, seven questions
Before I could investigate the discipline, I had to clear the ground. Because the reason planning has no canon is partly that the word “planning” has been used to mean at least seven different things, and the collapse prevents the diagnosis.
When a company says “our strategy failed,” at least seven different things might have happened:
The strategic choice might have been wrong: the firm picked the wrong market or position.
The forecast might have been wrong: the world shifted.
The targets might have been confused with commitments: aspirational numbers treated as if they were binding decisions.
The budget process might have starved the priority.
The functions might have been uncoordinated.
The plan might have been brittle, with no contingency for when assumptions broke.
Or execution might simply have been late and sloppy.
Seven pathologies, seven root causes, seven remedies. But if the organization collapses them all into one label, “the strategy failed,” it learns nothing. It cannot diagnose. It cannot correct. It restarts the cycle hoping the next strategy will magically implement itself.
The collapse happens because modern management language treats distinct disciplines as interchangeable:
Forecasts are beliefs. Plans are commitments. Budgets are authorizations. Targets are aspirations. Execution is work. Control is correction. Strategy is winning logic.
Those seven sentences should be memorized by every management team. Any organization that cannot maintain these distinctions will eventually make category errors serious enough to destroy value. A plan is not a strategy. But, also, a plan is not a forecast. A plan is not a budget. A plan is not a target. A plan is not a project schedule. Each is an input, a binding, or an artifact within planning. Not the discipline itself.
Because planning lost its independent identity, three inferior substitutes rushed in to fill the vacuum:
Planning as annual budgeting reduces commitment design to a finance ritual, run by FP&A on a calendar that bears no relation to strategic cycles.
Planning as project scheduling narrows enterprise-level resource synchronization to downstream task choreography handled by a PMO.
Planning as strategy theater decorates slides with the word “strategic” and calls it a day. Each captures a real fragment. None captures the discipline.
Once you accept that planning is distinct from all of those things, the real question opens up: what is it?
5. Governable commitment
Here is my positive definition.
Planning is the discipline by which an organization turns a chosen direction into funded, synchronized, contingent, and revisable commitments under explicit assumptions.
Every word is load-bearing.
“Funded” means real money and real people are attached.
“Synchronized” means commitments are designed to fit together across functions and time.
“Contingent” means the plan acknowledges uncertainty and contains branches.
“Revisable” means there is a mechanism for changing course before crisis forces it.
“Explicit assumptions” means the plan’s model of reality is visible, not buried.
But the critical move, the one that changes how one should read every planning document, is this: the unit of planning is commitment, not aspiration.
Most documents that call themselves “plans” are collections of aspirations. “Grow fifteen percent.” “Improve customer experience.” “Lead the category.” “Accelerate digital transformation.” These are aims. They may be reasonable aims. But they are not commitments.
A real commitment has six properties:
an owner who is personally accountable,
a resource implication that is actually budgeted,
a time implication specific enough to review against,
a dependency structure that identifies what else must happen for it to succeed,
a review rule that specifies when and how progress will be assessed,
and a stated consequence if conditions change.
Without those six properties, you do not have a plan. You have a wish list.
And as I flagged earlier when describing the typical planning deck: plans contain two fundamentally different kinds of objects, and most plans collapse them into one. Commitments are actions the organization can directly execute. Hypotheses are external responses the organization can influence but not control. A commitment can be completed by willpower, authority, and resources. A hypothesis can only be tested, and the test takes time the plan rarely accounts for.
Every major line item in a serious plan should be tagged: commitment or hypothesis. Without that distinction, the organization will eventually misread the completion of actions as proof of progress in outcomes, which is one of the most reliable ways to produce expensive failure.
The best criterion for judging planning quality is therefore not document elegance or analytical depth. It is governability:
Can the enterprise actually steer?
Are assumptions explicit?
Are commitments owned?
Are allocations real?
Are dependencies synchronized?
Are contingencies prepared?
Is reallocation authority live and fast enough to matter?
If yes, the organization is well-planned. Even if the presentation is ugly.
If no, the PowerPoint is wallpaper.
6. Planning lives in allocation
The commitment test tells you whether a plan is real. But there is a harder question: where does commitment actually become real in organizational life?
The answer does not live in any planning document.
It lives in allocation.
Joseph Bower’s 1970 study of corporate resource allocation, titled Managing the Resource Allocation Process: A Study of Corporate Planning and Investment, is perhaps one of the most important management books most strategists have never read. Bower spent years inside large corporations watching how investment decisions actually happened. What he found demolished the textbook model. Organizations do not allocate resources through serene analytical optimization. They allocate through a layered process:
definition: what gets proposed,
impetus: what gains momentum through middle-management championing,
and structural context: the rules, incentives, and structures that determine which proposals gain impetus and which die quietly.
The implication rewires how to think about strategy and planning. Bower shows that resource allocation drives realized strategy at least as much as strategy drives resource allocation. The actual plan of the organization is what it truly does, as opposed to what it says. And it is revealed by where money, people, time, and attention actually flow.
As Martin says, “there is no conflict between strategy and planning. Effective strategy needs thorough planning. And planning is of limited value without strategy.”
Aaron Wildavsky arrived at the same conclusion from the public sector. According to him, budgets are not neutral financial exercises. They are the central political process by which priorities become authoritative. The budget reveals what the organization actually values, regardless of what its mission statement proclaims.
Together they deliver the hardest sentence in the planning literature:
The real plan is the pattern of commitments the organization has actually funded, staffed, sequenced, protected, and agreed to review.
Not the presentation. Not the memo. Not the town hall speech. The allocation.
V. O. Key stated the foundational question: “On what basis shall it be decided to allocate X dollars to activity A instead of activity B?” That question is the heart of planning. Every enterprise faces it. Remarkably few have a principled answer beyond “last year’s baseline plus a political negotiation.”
This is also what Martin means when he observes that budgeting “enjoys possession” in organizations. The budget determines authority to spend. Strategy might be the source of wisdom, but budgeting is the source of motion. If a strategic priority is not baked into the budget, it will not happen, regardless of how much the CEO endorsed it at the offsite. Martin’s observation is precisely the Bower-Wildavsky insight dressed in executive language. What Martin points at from the strategy side, Bower and Wildavsky had mapped from the planning inside.
Clayton Christensen’s disruption theory, often taught as a strategy theory, is at its root an allocation theory. Christensen was Bower’s doctoral student and built directly on his framework. The reason incumbents get disrupted is not bad strategies; most incumbents have excellent strategies for their existing markets. The problem is that their resource-allocation process systematically starves unfamiliar opportunities, favoring proposals with high certainty, large addressable markets, and established customer demand. Precisely the attributes disruptive technologies lack. The innovator’s dilemma is an allocation dilemma. The planning system is structurally biased against the very investments the strategy would require if the firm took disruption seriously.
A plan without an allocative layer, without real money and real protection attached to its priorities, is fiction.
7. Uncertainty does not abolish planning
At this point, we meet the strongest objection from anti-planners.
“The world is volatile. Plans go stale the moment they are written. In a turbulent environment, better to stay agile and adapt.”
The strong version of this argument rests on a genuine insight: rigid blueprints under genuine uncertainty are dangerous. They create false confidence, resist revision, and punish adaptation.
But the mistake is in the inference. Rigid planning fails. That does not mean planning fails. It means a specific kind of planning fails: the attempt to specify a fixed course under conditions where key variables are unknown.
Good planning under uncertainty looks completely different from blueprinting.
And four examples, drawn from entirely different institutional worlds, whose practitioners mostly never read each other, converged on the same answer:
Pierre Wack, head of scenario planning at Royal Dutch Shell in the early 1970s, built the most celebrated case. When the 1973 OPEC production cut sent the global oil industry into crisis, Shell was the only major oil company with prepared contingency responses. Not because Wack predicted the crisis, he explicitly did not. But his scenario process had forced Shell’s managers to rehearse what they would do if oil-producing nations restricted supply. He called this “reperceiving”: developing what he termed “horse vision,” a peripheral awareness that lets you see the world differently before you are forced to. The scenarios changed mental models so managers could recognize the crisis and act while competitors were still trying to understand what had happened. Arie de Geus, Shell’s Group Planning Coordinator, distilled the lesson: “The purpose of planning is not to produce plans but to produce institutional learning.”
Dwight Eisenhower said something parallel in 1957: “Plans are worthless, but planning is everything.” The specific document will be overtaken. The process builds readiness and interpretive capacity that outlast it.
Helmuth von Moltke the Elder stated the principle a century earlier: “No plan of operations extends with any certainty beyond the first encounter with the main enemy forces.” His response was not to stop planning. It was to plan more deeply. Through branches (prepared alternatives for contingencies within the current operation), sequels (follow-on plans depending on how outcomes unfold), commander’s intent (a directional frame that allows subordinates to adapt without losing coherence), and decision points (explicit moments where incoming information should trigger a pre-prepared pivot). The military never assumed away opposition, delay, or friction. It treated these as the normal conditions of planning.
And from financial economics, Dixit and Pindyck’s real-options theory showed why staged commitment often beats all-in commitment. When investments are partly irreversible and uncertainty genuine, the planning question is not only “What do we want to do?” but “What optionality are we destroying by committing too early?” Under irreversibility, it can be optimal to wait until benefits are at least twice the investment cost before committing; a result that, translated into planning language, says “don’t pre-invest; pre-clear.”
Shell planners, a five-star general, a Prussian field marshal, and financial economists, none of them reading each other’s work all arrived at the same place. Uncertainty does not abolish planning. It changes what good planning looks like.
Which raises an obvious question. If these four examples from very different worlds surface the same insight, there may be a wider canon here. Not just scattered witnesses, but intellectual schools of planning, each with its own texts, thinkers, and methodology. I went looking for them.
8. Seven schools of good planning
I used Claude as a research partner to map the literature, find the thinkers I had missed, stress-test the framing, and surface the deep texts I had not read. It turned out to be well suited for the task, because the planning canon is genuinely scattered across disciplines that do not cite each other, and the work of tracing the connections is exactly the work that AI is now very good at helping with.
What came back surprised me.
The planning canon exists. It has seven distinct schools, developed across the twentieth century in institutional contexts so different that the thinkers mostly never talked to one another. Each school produced its own deep texts, its own methodology, and its own blind spots. Three of them will already be familiar from the uncertainty section above: the scenario school (where Wack and de Geus operated), military doctrine (Moltke and Eisenhower), and capital planning (the real-options lineage of Dixit, Pindyck, McDonald and Siegel). Those were three partial glimpses of a larger set. Here is the full picture.
Together these schools cover an enormous intellectual territory. Individually, none of them is the whole of planning, which is why no single one of them ever grew into a canonical center the way Porter became the canonical center of competitive strategy.
Here are the seven, in the order I now think of them.
Scenario planning is the Shell school. Herman Kahn (who originated the method at RAND), Pierre Wack and Ted Newland (who adapted it for corporate use at Shell), Arie de Geus (who crystallized its purpose), Kees van der Heijden (who turned it into “the strategic conversation”), and Peter Schwartz (who popularized it globally). The core insight is that planning’s output is not a document; it is changed minds. The planner’s job is to expand the mental models of decision-makers so they can recognize futures when they arrive, not to predict which future will come.
Resource allocation theory is the Bower-Burgelman school. Joseph Bower’s 1970 study and Robert Burgelman’s twelve-year study of Intel together demonstrated that organizations do not allocate resources rationally, they allocate through layered organizational processes that determine which proposals gain momentum and which die. The core insight: if intended strategy does not drive resource allocation, resource allocation drives realized strategy. The planning process is a strategy-deploying process, whether you admit it or not.
Budgeting theory is the public-finance lineage. V. O. Key’s 1940 question about why X dollars go to activity A rather than B, Aaron Wildavsky’s studies of the politics of the budgetary process, Allen Schick’s work on budget reform. The core insight: budgeting is not an administrative afterthought to planning; it is planning made concrete and politically authoritative. You cannot govern what you cannot budget.
Operations research is the quantitative backbone. Patrick Blackett, George Dantzig, and the post-war tradition that turned military logistics problems into mathematical optimization. Linear programming. Queueing theory. Inventory models. Routing algorithms. The core insight: many planning problems have formal structure that can be solved rather than deliberated, when the objective is clear and the constraints are explicit.
Capital planning is the finance lineage. Joel Dean’s foundational work, Stewart Myers’ development of real options thinking, and the broader tradition that made staged investment under uncertainty into a rigorous framework. The core insight: the value of flexibility, staged commitment, and the option to abandon. They transform planning from one-shot NPV calculation into a dynamic framework that explicitly values learning.
Forecasting science is what planning’s epistemic foundation actually looks like when done seriously. Scott Armstrong’s evidence-based forecasting principles, Spyros Makridakis and the M-competitions, Philip Tetlock’s Good Judgment Project demonstrating that forecasting skill is trainable and measurable. The core insight: the quality of any plan is bounded by the quality of the beliefs feeding it, and forecasting quality can be systematically measured and improved.
Military planning doctrine is the oldest and most institutionalized of the seven. Helmuth von Moltke the Elder’s mid-nineteenth-century reforms in the Prussian staff system, Dwight Eisenhower’s synthesis in modern joint operations, the current U.S. Joint Publication 5-0, and Stephen Bungay’s translation of the Prussian Auftragstaktik tradition for corporate practice in The Art of Action. The core insight: plans are worthless but planning is everything; the value lies in branches, sequels, commander’s intent, and the organizational capability to adapt, not in the document itself.
Above these seven schools sit two figures who do not fit cleanly into any of them but frame the entire field:
Herbert Simon is the meta-theorist of planning’s cognitive constraints: his work on bounded rationality tells you what planning cannot do, which is achieve comprehensive optimization across all alternatives. And, therefore, what it must do instead, which is satisfice, decompose, and alternate between architectural redesign and incremental adjustment.
Russell Ackoff is the field’s deepest philosopher, the Peter Drucker of planning, the author of Creating the Corporate Future and, unfortunately, the only major thinker who placed strategy within planning rather than separate from it. Ackoff is philosophically understandable but operationally misleading; in the managerial sequence, strategic direction logically precedes the commitment design planning performs. However, his broader point matters: planning is itself a design discipline of the highest intellectual order, not a clerical function downstream from the “real” thinking.
Seven schools. Two framing thinkers above them. And here is what made the picture start to click for me: as I mapped what each school actually solved for, I noticed that two of them are not optional.
Bower’s resource allocation and Wildavsky’s budgeting are the physics of organizational life. Every strategy, from every school, must eventually become a flow of money and attention toward some activities and away from others. There is no version of “translating strategy into action” that does not pass through resource allocation and budgets. Bower tells you how resources actually flow. Wildavsky tells you how budgets actually get made. Skip either one and you are planning in a fantasy world where resources allocate themselves rationally and budgets reflect pure intent. Every other planning school presumes these two; none of them substitutes for them.
The other five form the variable ring around that core. Which ones you emphasize depends on your dominant strategy type and your uncertainty profile. But resource allocation and budgeting are always there, underneath, whether the organization acknowledges them or not.
This observation turns the scattered canon into something that looks like architecture. Because if two schools are mandatory and five are conditional, then the next question is: what structural problem is each school actually solving?
That question cracked the puzzle open. The seven schools are not seven different disciplines. They are seven intellectual traditions converging on four structural problems that every serious plan must solve. Some problems have more than one witness school: resource allocation and budgeting both work on where commitment becomes real; scenario planning and forecasting science both work on what the plan believes about the world; military doctrine and real-options thinking both work on how the plan revises when conditions change. Operations research cuts across several. That is why the counts differ: seven intellectual traditions mapping onto four structural jobs a plan must do.
Here is the architecture.
9. The four-layer architecture
Every serious plan has four irreducible layers. Any plan missing one may exist as a document, but it is structurally incomplete as a planning system. Here are the four layers, and with each, the diagnostic question that tells you whether it is present in your organization.
Layer 1: The assumption layer
What does the plan believe about the world?
Every plan embeds a model of reality: assumptions about demand growth, cost trajectories, competitor behavior, regulatory shifts, technology readiness, macroeconomic conditions. Those assumptions always exist. The question is whether they are visible or buried.
When this layer is done well, the plan’s key assumptions are stated explicitly, ranked by impact and vulnerability, and connected to signposts: observable indicators that tell the organization whether the assumption is holding or breaking. This is what Lafley, Martin, Rivkin, and Siggelkow mean when they argue that strategy should specify “what would have to be true” and identify the least likely, most critical conditions for success. It is what Rita McGrath and Ian MacMillan formalized as discovery-driven planning: start with the reverse income statement and use milestones to surface assumptions before full commitment. It is what James Dewar built into assumption-based planning at RAND: vulnerable crucial assumptions become monitored objects with pre-designed responses.
The plan contains not a single forecast but a scenario structure: a base case and at least two alternative cases representing materially different futures. The organization knows, before those futures arrive, what it would do differently in each one.
This is also where the base-case-plus-two-costumes pattern I described earlier is most clearly a failure. Pierre Wack named the failure mode decades ago, calling these “first-generation scenarios”: mere quantification of obvious uncertainties around a single worldview. Martin puts it more directly: scenario analysis that produces no prescriptive “so what” is analysis for analysis’s sake. The decision-switch test, drawn from this tradition, is unforgiving: if a scenario does not change the identity of the first binding constraint, the chosen leading indicators, or the actions you would take, it is not a scenario for planning purposes. It is a range.
Real scenario planning presents fundamentally different causal worlds, not variations on one theme. A real scenario has different assumptions, a different first binding constraint, different leading indicators, different tripwire thresholds, and a different action package. It is not “what if revenue is higher or lower?” It is “what if the customer problem we are solving stops being the problem customers have?” Or: “what if the channel we depend on restructures around a different economic model?” Or: “what if the regulatory posture shifts from permissive to restrictive in our primary geography?” Each of these requires a different plan, not a different number.
The broader forecasting literature reinforces the discipline. Armstrong’s evidence-based forecasting principles demonstrate that simple models often outperform complex ones, that combining methods beats individual forecasters, and that calibrated uncertainty ranges beat false-precision point estimates. Tetlock’s Good Judgment Project demonstrated that forecasting skill is trainable. The assumption layer can be made systematically better.
When this layer is absent, the plan quietly assumes certainty. There is no scenario structure beyond three drivers and one worldview. No signposts. No visible premises.
The diagnostic question: Can your leadership team name the key assumptions that, if wrong, would invalidate the current plan? And can they name the signposts that would tell them those assumptions are breaking? If the answer takes more than two minutes, the assumption layer may be missing.
Layer 2: The allocative layer
What gets funded, staffed, and protected?
This is where planning becomes materially real. An unfunded priority is a fantasy, regardless of how passionately the CEO endorsed it at the offsite.
When this layer is done well, the budget reflects the stated priorities. There is a visible, principled process for deciding why activity A receives resources instead of activity B, the foundational question. Strategic priorities receive protected funding that cannot be raided by operational urgencies. Targets, forecasts, and resource allocations are treated as three separate instruments rather than collapsed into one budget number, a distinction the Beyond Budgeting movement has made central to its operating philosophy.
The canonical case is Svenska Handelsbanken, the Swedish bank that operated for over forty years without annual budgets, consistently outperforming its Nordic banking peers. Handelsbanken separated target-setting (relative, not absolute), forecasting (rolling, not annual), and resource allocation (decentralized to branches, not centrally rationed). The result was a planning system that could adapt continuously without the gaming and political negotiation that annual budgets produce. Bjarte Bogsnes and the Beyond Budgeting Round Table have documented this case in detail.
When this layer is absent, the budget is set through incremental negotiation (last year’s number adjusted by political pressure), and bears little relationship to the stated strategy. The allocation process is structurally decoupled from the strategy process. They live in different calendars, different meetings, different organizational functions, and different incentive systems. The strategy says one thing, and the money says another. The money wins.
The diagnostic question: Lay your strategy deck next to your budget. If an intelligent outsider who had never met your team examined both documents, would they say they were about the same company?
Layer 3: The coordinative layer
How do commitments fit together across people, functions, and time?
A firm can have clear assumptions and fully funded priorities and still fail because the commitments do not interlock. Marketing plans a product launch for March. Operations cannot ramp production until June. Two product teams both need the same data-engineering squad in Q2 and neither knows it. Finance approved the headcount but the recruiting pipeline was not informed. Sales promised a feature that Engineering has not prioritized. The pieces never fit.
When this layer is done well, the organization operates from one coherent, reconciled set of commitments. Cross-functional dependencies are mapped and resolved before they become execution crises. Vertical alignment runs from corporate direction down to team-level objectives. Horizontal alignment ensures that functions are not planning in isolation and discovering conflicts only during execution.
Two traditions are the primary witnesses. Hoshin kanri, the Japanese policy-deployment system formalized by Yoji Akao, was built to solve exactly this problem. Its signature mechanism, “catchball,” involves structured back-and-forth negotiation between organizational levels. The senior team sets direction. The next level proposes how to achieve it and what resources it would require. The senior team reviews, challenges, and adjusts. The next level responds, refines, and commits. The process cascades downward and laterally until commitments are reconciled both vertically and horizontally.
Critically, catchball is not a one-time cascade. It is a disciplined iterative negotiation that surfaces conflicts and dependencies early, when they can be resolved in conversation rather than late, when they have already become crises in execution.
Integrated business planning, the modern descendant of sales-and-operations planning, attacks the same coordination problem from the supply-chain side. The core discipline is a monthly cycle that reconciles demand, supply, finance, and portfolio into one synchronized operating plan. When IBP works, Marketing’s demand plan and Operations’ supply plan and Finance’s financial plan are the same plan, reviewed and reconciled every month; not three separate plans that discover their contradictions in the quarterly review.
When the coordinative layer is absent, functions plan independently and discover conflicts in execution. The coordination failure is routinely misdiagnosed as an execution failure, because by the time it becomes visible, the damage looks like missed deadlines rather than what it actually is: a planning system that never synchronized the commitments in the first place.
The diagnostic question: If you asked Marketing, Operations, Finance, and Engineering each to describe “the plan,” would they describe the same object? Not the strategy, but the specific set of commitments with the same timelines, dependencies, and resource assumptions?
Layer 4: The adaptive layer — ceiling and floor
How does the plan change when conditions change?
This is the layer that separates planning from blueprinting. A plan that cannot revise itself when assumptions break is not a plan; it is a prayer. And this is where the practice of most organizations is thinnest.
The adaptive layer has two sub-disciplines that must be designed separately, because they respond to different kinds of deviation. I will call them ceiling discipline and floor discipline.
Ceiling discipline addresses the question: if things go materially better than expected, what is the first crux that will cap performance, and when must we act to relieve it? Upside surprises can destroy value almost as effectively as downside surprises. A product that wins demand but cannot fulfill it destroys trust. A channel that scales faster than service capacity degrades experience. A business that sells ahead of working capital suffocates itself. A firm that finally finds traction but cannot hire or certify people fast enough hands the market to imitators.
Rumelt’s concept of the crux from is the correct master frame here. The crux is “the key issue where action will best pay off.” the decisive difficulty on the path of progress. Rumelt intended this for strategic diagnosis, but it applies with equal force to upside planning. Under an extraordinary success scenario, what is the first element whose capacity or feasibility limit gets reached with no realistic short-term substitute? Production capacity. Talent pipeline. Regulatory approval queue. Customer support load. Channel shelf space. Management bandwidth. Working capital. Each of these can be a ceiling, and the first one to bind is the one that matters.
Ceiling discipline does not mean pre-investing for success that has not yet happened. That is one of the more expensive forms of bad planning, and real-options theory explains why: when investment is irreversible and uncertainty is high, the value of waiting is often large. Ceiling discipline means something more precise: pre-clear, don’t pre-invest. Identify the likely first crux. Estimate the lead time to relieve it: how long it takes from “decide to expand” to “expansion actually in place.” Identify the signposts that will indicate you are heading toward the crux. Pre-authorize the relief play. And calculate the trigger point using a simple formula:
Trigger point = time-to-hit-crux − lead time to relieve crux − safety margin
If the trigger point is negative, you have already waited too long. Before the trigger, spend only on low-regret enabling moves: permits, vendor qualification, modular designs, cross-training, contract options, financing lines. These are cheap option premiums. After the trigger, exercise the option. Before it, keep it open.
Floor discipline addresses the opposite question: if things go materially worse than expected, what will tell us early, what do we pull back first, and what loss boundary are we refusing to cross?
The weakest possible approach to downside is the ritual “bad case” deck that lowers top-line assumptions and stops thinking. Serious floor discipline is organized around a tight set of design questions, drawn from the convergence of stress-testing doctrine, discovery-driven planning, and real options: What would tell us early that the core assumptions are failing? What threshold actually matters? What action do we take immediately when that threshold is breached? What loss are we explicitly unwilling to exceed?
A signpost is an observable indicator that reveals which path reality is taking. A tripwire is the threshold on that indicator that triggers a predefined action. Signposts without tripwires are monitoring. Tripwires without predefined actions are alarmism. A tripwire that merely triggers “further analysis” is usually too weak. The recovery-planning frameworks in banking, which exist because regulators saw what happens when monitoring is not coupled to pre-committed action, are instructive: thresholds are designed to activate escalation processes and recovery options early enough to be effective.
Most downside plans fail because they imagine only two states: continue or total failure. This is far too coarse. The correct design is a pullback ladder: slow discretionary spend, pause hiring, narrow scope, delay expansion, renegotiate commitments, shift to pilot mode, exit. The crucial question is: what gets pulled back first? If the answer is not written down before stress emerges, the organization will improvise under pressure and usually cut the wrong things. Brand marketing gets cut because it has no defender in the room. R&D gets cut because its return is long. Training gets cut because it is discretionary. The cuts that feel easy in the moment are often the ones that destroy long-term position.
Finally, floor discipline requires an explicit maximum tolerable loss. A practical formulation:
Maximum Tolerable Loss = sunk cost to next decision gate + shutdown cost + committed liabilities + working-capital unwind + material tail provisions
This is a disciplined accounting of exposure. It forces the organization to state, before stress arrives, what it is actually willing to lose. Without this number, losses tend to creep past any threshold that would have triggered a stop, because each incremental commitment feels small relative to sunk cost.
One more nuance matters. Some floors must be structural, not trigger-based. Cyber catastrophes, liquidity freezes, catastrophic quality failures, hard covenant breaches can move faster than detection and response. No tripwire helps after the cliff edge. For these, the planning response is ex ante structural protection: lower leverage, more liquidity, redundancy, segregation of duties, kill switches, exposure limits, insurance.
Where the hazard can jump faster than the organization can react, resilience must be designed in, not monitored.
Military doctrine, from Moltke through modern joint planning and Stephen Bungay’s The Art of Action, is the primary synthesis tradition for the adaptive layer. Bungay identifies three gaps every organization faces: the knowledge gap (we cannot know everything about the situation), the alignment gap (people will not always do what we plan), and the effects gap (actions will not always produce the outcomes we intended). The natural but wrong response is tighter control. The right response is clearer intent, more delegation of method, and pre-built adaptation mechanisms.
The diagnostic questions for the adaptive layer are two. If the answer to either is no, the full adaptive layer is not there.
For ceiling: have you identified the first binding crux under success, the lead time to relieve it, and the trigger point? Are pre-approved option premiums in place?
For floor: do you have named signposts with tripwires that trigger pre-authorized actions, a pullback ladder specifying what gets cut first, an explicit maximum tolerable loss, and structural protections for fast-moving hazards?
10. Most organizations fail three of four
Here is what my personal audit reveals, applied across the organizations I have worked with and studied.
Most companies pass on layer three. The annual planning cycle, whatever its flaws, forces cross-functional conversation. People sit in rooms together. They present their numbers. Some synchronization happens, even if imperfect. The coordinative layer is usually present, even if it is not strong.
Most companies fail on the layers one, two and four.
The assumption layer is typically absent. Or, worse, replaced by the three-costume pattern that performs uncertainty management without delivering any of it. Nobody can name the two or three assumptions that would invalidate the plan. The plan pretends certainty it does not possess.
The allocative layer is typically disconnected from strategy. The budget reproduces last year’s pattern regardless of what the strategy says. Strategic priorities get seed funding in September and see it quietly reallocated to operational fires by February.
The adaptive layer is typically nonexistent on both sides. Ceiling discipline does not exist; nobody has named the first binding crux of success or calculated a trigger point. Floor discipline does not exist; there are no signposts with real tripwires, no pullback ladder, no maximum tolerable loss, no structural protections against fast-moving hazards.
The result is a characteristic organizational pathology: synchronized action toward goals nobody examined carefully, funded through political negotiation nobody questioned, with no mechanism for detecting when assumptions break, no preparation for the upside crux that will appear if things go well, and no architecture for the floor actions that will be needed if things go badly.
The typical plan coordinates. It does not govern.
Governability, the capacity to actually steer, requires all four layers. When all four are present, something remarkable happens. Assumptions are visible and revisable, so the organization can detect environmental shifts before they become crises. Resources actually move to match priorities, so the strategy’s financial expression matches its competitive intent. Functions operate from one coherent set of commitments, so execution is synchronized by design rather than by improvisation. And reallocation happens before crisis forces it: upside options are pre-cleared, downside tripwires are pre-authorized, and the organization adapts continuously rather than in panicked lurches.
These four layers contain inherent tensions that cannot be resolved, only managed. The planning system must be centralized enough to cohere but decentralized enough to learn. It must be stable enough to commit but flexible enough to revise. It must be disciplined enough to govern but light enough to avoid bureaucratic calcification. These are not incidental nuisances. They are the permanent paradoxes of the discipline. Anyone promising a planning system that eliminates these tensions is selling a fantasy. Good planning lives inside the tensions.
11. Why the canon was never assembled
The four-layer architecture also resolves a puzzle that had nagged me: why no single great theorist emerged the way Porter emerged for competitive strategy.
The answer is not that planning lacks intellectual substance. The answer is that planning is a mixed discipline, with four irreducible dimensions, and the seven schools each solved for a different combination. The scenario school solved the assumption layer. The resource-allocation school and the budgeting school solved the allocative layer. Hoshin kanri and IBP solved the coordinative layer. Military doctrine and real-options thinking solved the adaptive layer.
Each tradition developed its own texts, methodologies, and thought leaders. None of them produced a synthesis that treated planning as a single discipline with a coherent architecture, because no one of them had the whole architecture in view.
The canon is federated because the discipline is mixed. But the material exists. Ackoff as philosopher. Simon as cognitive foundation. Bower and Wildavsky as irreducible core. Wack and de Geus on perception and learning. Armstrong and Tetlock on epistemic discipline. Moltke and Bungay on contingency architecture. Myers on staged commitment. Akao on deployment. This is the canon. It has been there all along. It simply never got assembled into one field with one name.
12. Strategy chooses, planning commits
I am a strategist. I believe doing strategy well: making genuinely distinctive choices, building activity systems that produce real competitive advantages, choosing trade-offs that competitors cannot or will not match. I have seen so few organizations that can truly do strategy, genuinely make strategic choices rather than produce planning documents dressed in strategy language, that I remain convinced strategy is a most critical problem. Martin is right about the colonization. Rumelt is right about the dismissive attitude toward strategy that pervades corporate practice.
But there is a second discipline, downstream in logic and perhaps equal in importance, that determines whether good strategy can exist in reality. That discipline is planning. It has an object: governable commitment. It has an architecture: assumption, allocative, coordinative, adaptive. It has a canon: seven schools, two of them irreducible, with Ackoff as philosopher and Simon as foundation. And it has a diagnostic: four questions, one per layer, the adaptive one asked in two halves, that take ten minutes to ask and tell you more about the health of an organization’s planning system than a hundred pages of strategy decks.
Step one: Unmix strategy and planning. Understand that they answer different questions.
Step two: Do great strategy. Make genuinely distinctive choices. This is the hardest step.
Step three: Do great planning. Convert those choices into funded, synchronized, contingent, and revisable commitments, with assumptions made explicit, allocation that matches intent, coordination that actually synchronizes, and an adaptive layer that prepares for both the first binding crux of success and the first early signal of failure.
Frankly, almost nobody does any of these steps well. And step three is where the literature has least been properly canonized.
Strategy without planning is a speech delivered to an empty room. Planning without strategy is coordinated drift toward nowhere in particular. Strategy is an art and a practice. Planning is a discipline. Both deserve canons. Both deserve the highest quality of intellectual attention a leadership team can bring to bear.
Strategy chooses a direction. Planning decides whether that choice can survive contact with money, capacity, time, and surprise. The organization does not become what it declares. It becomes what it actually funds, sequences, protects, and revises.
The shelf in the bookstore is still empty. Perhaps it’s time to start filling it.
Thank you for this. I've been on a journey to understand strategy & developed my own frame to link it to my experience as a finance professional. Linking great thoughts from Roger or Rumelt to commitment, resource allocation, budgeting etc gave me the link I really needed to take my own journey forward.
The link from strategy to planning & then breaking planning into different buckets makes it so obvious in hindsight after reading your work but very easy to miss. I've seen budgeting as last year plus a percentage with scenario analysis meaning only three scenarios with different assumptions on the numbers without changing the underlying model assumptions.
I took weeks to read through as I as digesting each section before moving to the next.
Keep writing !! and thank you!!
The intellectual gap on this discussion is closed by Alejandro Salazar in “Emergent Strategy and the Death of the strategic plan”
You should definitely take a look