DEEP READ · 926
Good Strategy Bad Strategy: The Difference and Why It Matters · Richard Rumelt · 2011
That document your company calls its strategy—the one promising to become the industry leader, hit double-digit growth and put the customer first—is not a strategy; it is a list of wishes. Real strategy is a cold-blooded craft: say plainly what is blocking you, choose one route around the obstacle, then put your limited resources behind that route and cut the rest. The book's great contribution is not teaching you to write good strategy but handing you a detector for waffle—because bad strategy is not the residue of incompetence. It is a recognisable way of writing and thinking, with a fixed set of symptoms.
Richard Rumelt teaches strategy at UCLA's Anderson School and is widely described as the strategist's strategist; concepts he developed in the 1980s, such as isolating mechanisms, are part of the foundation of the theory of why some firms keep an advantage for decades. This 2011 book is the ledger of forty years of teaching and consulting, written for practitioners who have already sat through too many hollow strategy offsites. What sets it apart from its shelf-mates is that it spends a full third of its length on how to recognise the bad—and that half is where its real force lies.
Rumelt strips strategy down to three components, and any strategy—corporate, military or personal—needs all three at once:
Steve Jobs' 1997 return to Apple is the clearest worked example. Diagnosis: the product line is far too long and too tangled, the company is bleeding into complexity of its own making, and it has no prospect of beating the Wintel camp head-on in mainstream PCs. Guiding policy: survive, cut to the bone, hold a defensible patch of ground on design and experience, and wait for the next wave. Coherent actions: kill nearly every model, keeping one machine in each cell of a consumer/professional by desktop/portable grid; cut distributors, inventory and engineering projects; settle with Microsoft to secure Office for the Mac. Rumelt records asking Jobs in 1998 what his long-run plan was and being told, in effect, "I am going to wait for the next big thing." That sounds like the absence of strategy and is in fact the strategy: survive, focus, and hold your fire until a wave arrives that your strengths can ride. (The wave turned out to be the iPod.)
| Kernel element | Question it answers | The usual counterfeit |
| Diagnosis | What is actually blocking us? What kind of problem is this? | A flat list of "challenges and opportunities" naming no crux |
| Guiding policy | How do we get around it? What do we give up? | "Be the leader," "customer first"—ruling nothing out |
| Coherent actions | Who does what on Monday? Where do resources move? | Departmental wish lists that don't mesh, or cancel out |
Fastest test of any "strategy": sort it into these three boxes and see which one is empty.
This is the book's most underrated element. Rumelt keeps returning to the physician's analogy: a patient saying "my back hurts" has reported a symptom, not a diagnosis; the doctor's job is to compress scattered information into a judgement that can guide treatment. A good diagnosis substitutes for complexity—it deletes most of the facts, keeps the few that are doing the work, and by doing so points at action.
His Wal-Mart classroom case makes the point stick. Students usually explain the defeat of Kmart like this: Wal-Mart went to small towns, opened a big store, and the town could support only one, so it enjoyed a local monopoly. Rumelt's response is that the diagnosis picks the wrong unit of analysis. If a single store were the business, Wal-Mart's stores were too small and too remote to work. The real unit is not the store but a distribution centre plus the roughly one hundred and fifty stores packed around it—stores clustered within a truck's easy reach, replenishment and inventory run across the whole web. Change "store" to "network" and every odd-looking choice—the remote towns, the density, the owned logistics, the shared information system—snaps into a single design. That is what diagnosis does: not more data, but a different unit, so that loose facts suddenly assemble into a structure.
This is the most useful and most caustic part of the book. Bad strategy is not a competence problem; it is a genre you can learn to recognise:
| Hallmark | What it looks like |
| Fluff | Restating the obvious in esoteric vocabulary to manufacture depth |
| Failure to face the challenge | Never says what the obstacle is—so no proposal can be judged good or bad |
| Mistaking goals for strategy | "Double revenue in three years" is a desired result, not a way to get it |
| Bad strategic objectives | Either a dog's dinner of twenty co-equal items, or blue-sky castles in the air |
Hit any one of the four and the document can go back where it came from.
Fluff is the easiest to catch. Rumelt quotes a bank whose "fundamental strategy" was customer-centric intermediation. Translated: we take deposits and make loans. That is the definition of a bank, not the strategy of any particular one. Whenever an industry's standard operating procedure is dressed in abstract nouns and presented as strategy, this is what you are looking at.
Failure to face the challenge is the fatal one. The whole point of a strategy is to overcome some specific obstacle; if a document never states the hardest thing standing in the way, it cannot be evaluated at all, because with no problem defined, no proposal can be wrong. Rumelt's example is the 1979 improvement plan of International Harvester, then a giant of American farm and truck equipment: the document set out a string of objectives—better margins, more share—while never confronting the company's actual knot, its rigid labour relations and inefficient production system. The long strike that followed gutted the firm, and it never recovered. However elegant the goals, going around the crux amounts to zero.
Mistaking goals for strategy is the most common. "Grow 20% next year" is a goal. "Because our rivals are slow to respond, we are moving everything into a service network that shows up within 24 hours, and giving up price-shopping customers to pay for it" is a strategy. One-line test: if the passage never says what you will therefore stop doing, it is not a strategy.
Rumelt's explanation goes well past "they were stupid." First, the unwillingness or inability to choose. Strategy takes resources away from some people and hands them to others, which is a real conflict inside an organisation; the path of least resistance is to do a bit of everything and write the strategy as a list in which everyone has a share. It reads like consensus. It is an abdication of command.
Second, template-style strategy. The fill-in-the-blanks scaffold—vision, mission, values, strategy—lets anybody produce a respectable-looking document in an hour. What the template eliminates is precisely the hardest and most valuable part: diagnosis and trade-off. Fill in the blanks and the painful thinking has been perfectly skipped.
Third, a misapplied brand of positive thinking. Rumelt traces it to the nineteenth-century New Thought movement—an American religious-psychological current holding that thought itself creates reality—which flowed into business through a century of success literature. Its modern face is the belief that if the goal is bold enough and everyone believes hard enough, a way will appear. So "set an inspiring vision" gets mistaken for the work of strategy. He does not deny that motivation and confidence matter; he objects to substituting exhortation for analysis. A rousing speech cannot answer the question of why we, specifically, should win.
The most basic idea in strategy, Rumelt says, is the application of strength against weakness. An organisation's force is always limited, so the question is never "how do we try harder" but where the same effort produces a disproportionate result. He breaks leverage into three parts, all required:
One of his best illustrations is the 1991 Gulf War. Commentary at the time treated the strategy as overwhelming force. What actually did the work was design: a frontal feint pinned the Iraqi army along the Kuwaiti line while the main armoured force swung hundreds of kilometres west through open desert and came in behind—the "left hook." Force is not strategy; the arrangement and concentration of force is. Read the other way, this is also why the weaker side can win: a challenger far outmatched needs only to find the fragile link that the opponent's strength depends on—the giant's crushing close-quarters power depends on your having to come close—and a very small force will do.
This is the most immediately usable idea in the book. One of a leader's most important jobs is translating a vague, paralysing situation into an objective close enough to be genuinely achievable. A proximate objective is one that already lies within the organisation's reach—still a stretch, but not a castle in the air.
Rumelt uses the 1961 American commitment to the Moon. "Land a man on the Moon and return him safely before this decade is out" was good strategy rather than fluff because the state of the art—rocket thrust, orbital mechanics—had been pushed far enough that feasibility was evidence-backed. The goal was high but reachable, and utterly unambiguous: a date, an action, and a pass/fail test. Contrast "achieve leadership in space," which has no boundary, cannot be verified, and tells nobody what to do tomorrow.
The mechanism underneath is that ambiguity eats an organisation's energy. A vague objective lets every department interpret and act for itself, and the forces cancel; a sharp proximate objective narrows uncertainty to the point where people can actually coordinate. So when facing something huge and chaotic, the right move is not to shout it louder but to cut it down to the layer where the next step can really be finished.
Some systems are governed by their weakest link rather than by the sum of their parts—as a chain is only as strong as its worst segment. Rumelt calls these chain-link systems, and they explain two things that look contradictory.
First, why partial improvement so often pays nothing. In a chain-link system, upgrading one link does not lift the whole: world-class manufacturing with mediocre design and a chaotic channel still yields a product nobody buys. Returns therefore arrive flat-then-jump—long investment with nothing visible, until every deficient link has been fixed and the gain lands all at once. It also explains why so many turnarounds die halfway: patience runs out before the payoff arrives, and on the books the effort looks like money spent for no effect.
Second, why certain advantages are nearly impossible to copy. IKEA is Rumelt's classic case: in-house design, flat-pack, huge out-of-town stores, customers doing the transport and assembly, a supply chain built for very low cost, restaurants and childcare. These are not a pile of separate good ideas but one interlocked chain. Copying any single link does nothing—adopt flat-pack without changing your stores and supply chain and your costs go up—while copying all of them at once means demolishing and rebuilding your own company. That is the sturdiest kind of moat: not a secret, but a design that only works whole.
The hard word in "coherent action" is not action but coherent. Rumelt separates incremental strategy—working roughly with the resources you have, improving in steps—from design-type strategy, where resources and moves are fitted tightly into a whole whose parts only work together: Hannibal at Cannae deliberately letting his centre fall back while the wings closed, each step alarming on its own and lethal in combination.
He draws a genuinely useful trade-off: the tighter the fit, the greater the power—and the more brittle, the harder to adjust, the more dependent on unified command. Design-type strategy therefore belongs to situations of relative resource disadvantage, where only ingenious configuration can win; with overwhelming resources there is no need to tie yourself down that hard. In other words, coordination has a price: it buys power and spends flexibility, and pursuing synergy for its own sake is itself a species of bad strategy.
Growth is the outcome of a good strategy, not a strategy. Rumelt is unsparing about growth-as-strategy: piling up revenue through serial acquisition usually creates no value, because the revenue was purchased with money while the organisational capability that would sustain the growth cannot be. Real growth comes from actually being better at something, so that demand arrives.
What stands in the way of change is two things. Inertia comes in three kinds: the inertia of routine—processes, metrics and habits have hardened, and people are not unwilling so much as muscle-trained; the inertia of culture—stubbornness at the level of values and identity, the slowest to shift; and inertia by proxy—the company is not the one refusing to change; its existing customers, channels and shareholders are, and it is locked in by its own relationships. Entropy is more mundane: with nobody pushing, an organisation drifts toward disorder—the product line sprawls on its own, focus dissolves on its own, standards slacken on their own. Holding a clear strategy is therefore not a decision but a job requiring continuous work.
These two have an inverted use, which is the book's most elegant turn: your rival's inertia is your opportunity. You usually win not because you are cleverer but because the incumbent knows it should change and cannot. Nvidia's defeat of 3dfx carries that signature: Nvidia ran overlapping design teams to compress its release cycle to a pace competitors could not match, and moved with the industry's common graphics interface standard, while the then-leading 3dfx clung to its proprietary interface and its old model, fell behind the rhythm, never caught up, and ended in bankruptcy and acquisition. Speed is not merely diligence; it is a weapon designed to burn an opponent's inertia.
The last layer is what separates this book from ordinary management writing: a good strategy is fundamentally a hypothesis about what will work. It rests on a judgement about the situation, and judgements can be wrong. So strategy work should look like scientific work: start from an anomaly—something that ought not to be the case, which everyone else explains away and you insist on chasing—form a hypothesis, then design an affordable way to test it and revise on the feedback.
From this comes his advice to the individual: independent judgement is the real source of strategic capability. The commonest organisational failure is not miscalculation but a whole room hypnotised by the same consensus—because everybody is doing it, nobody asks why it should work. His remedies are plain: write your own judgement down before the meeting (only a written judgement can be checked later), actively construct the strongest case against your own conclusion, and keep someone in the room whose job is to argue the other side. Strategy in the end is not a document but the habit of stating a judgement that could be proved wrong.
The book's three parts run demolish, build, train:
What it establishes reduces to one sentence: strategy is scarce because it requires saying out loud where you are weak and deciding what not to do—and both are deeply unwelcome inside organisations. Which is why most strategy sessions end up as goal-announcement sessions.
1. Bad strategy is not the absence of strategy but a genre with fixed moves—fluff, failing to face the challenge, mistaking goals for strategy, bad objectives. Learning to spot it pays faster than learning to write strategy.
2. Good strategy has exactly three parts: diagnosis (what is actually blocking us), guiding policy (how we will get around it and what we give up), coherent action (who does what on Monday). Sort any "strategy" into those boxes and the missing one shows instantly.
3. The power of diagnosis comes not from more data but from changing the unit of analysis: Wal-Mart did not win store against store, but as a distribution centre surrounded by a hundred and fifty stores—change the unit and loose facts lock into a structure.
4. The most basic idea in strategy is strength against weakness: force is finite, so what matters is not effort but finding the pivot where the same push moves the whole system, then loading it hard.
5. "Double in three years" is a goal, not a strategy. The one-line test is whether the passage says what you will therefore stop doing. No trade-off, no strategy.
6. One of a leader's highest-value acts is translating a huge chaotic situation into a proximate objective—already within reach, a genuine stretch, and sharp enough that nobody can fudge whether it was met. Ambiguity consumes an organisation's energy.
7. Some systems are decided by their weakest link: that explains both why single-point improvements pay nothing for a long time (the return lands only once every deficient link is fixed) and why interlocked systems like IKEA cannot be copied—one link is useless, all of them means rebuilding yourself.
8. The tighter the parts mesh, the more powerful and the more brittle: coordination has a price. Bind yourself into a precision machine when resources are short, not when you already hold the advantage.
9. Growth is the result of good strategy, not a strategy; what blocks change is inertia (of routine, of culture, and by proxy through existing customers and channels) and entropy (drift when nobody pushes)—and read backwards, a rival's inertia is your best opportunity, which makes speed a weapon rather than mere diligence.
10. Finally: a good strategy is a hypothesis that could be proved wrong, and the nerve to state one is itself scarce. The commonest failure in organisations is not miscalculation but a whole room hypnotised by consensus—so write your judgement down, build the strongest case against it, and never let "everyone is doing this" stand in for "here is why we win."