Good Strategy Bad Strategy: The Difference and Why It Matters

Miles: The twenty first of October, eighteen oh five, off Cape Trafalgar. Twenty seven British ships of the line against thirty three French and Spanish ones, and the custom is to form a line alongside the enemy's and trade broadsides.

Miles: Nelson won't form a line. He splits his fleet into two columns and sends them straight at the enemy, at right angles, so his leading ships take fire all the way in before they can properly answer it.

Miles: He's betting on two things he's noticed. The gunners on the other side are less practised than his, and the sea that day is heavy. By the day's end the combined fleet is broken, Britain hasn't lost a ship, and Nelson is dead.

Miles: Today: Good Strategy Bad Strategy, Richard Rumelt's book from twenty eleven. In two sentences: most of what gets called strategy is goals, slogans and wish lists, and the real thing starts with a plain account of what's in the way. And for anybody who's never been able to run on hints and atmosphere, a book that insists the problem be written down first is better news than it sounds.

Miles: Sixty Books. Presented by Stoar Consulting and Coaching.

Miles: Book thirty six, and a fresh question on the table. Miles here, and this is Sixty Books. Same deal: sixty management books, the best a century produced, taken one per episode, and each of them asked two things. What does it give somebody who leads, or works for, a mind that doesn't take the standard route. And how much of it still stands now that AI does part of the job. Nora brings the research, and she brought a stack.

Miles: Nora, last time closed on four numbers that tell a team how fast it moves and how quickly it recovers. None of them says whether the road goes anywhere worth going.

Nora: And this book would call that the normal condition. Rumelt's long complaint is that what passes for strategy in most organizations is a target with some encouragement attached.

Miles: So we've swapped the stopwatch for something else.

Nora: For a question you ask before anybody moves: what's wrong here. Five books of measuring, and this one opens by asking what the measuring was for. It's also the most confident of the next six, all about deciding with half the facts, so read it slowly.

Miles: Trafalgar is his introduction. What's in chapter one?

Nora: Apple, nineteen ninety seven. Steve Jobs is back, the company is losing money, and Rumelt's account of what he did is mostly subtraction. Fifteen desktop models cut to one. Every portable and handheld cut to one laptop. Printers and peripherals dropped completely.

Miles: And a big announcement to go with it.

Nora: None, and that's the detail he wants you to see. No revenue target, no profit target, no sermon about the future. Then in nineteen ninety eight Rumelt, consulting for an Italian telecom company at the time, gets to ask Jobs what the longer plan is. Jobs smiles and tells him he's going to wait for the next big thing.

Miles: Which sounds like no plan.

Nora: Which is why it's in the book. The cuts were the plan. Too many products and too little cash, named plainly, then a short set of moves all pushing one way. The waiting belonged to it. It wasn't a gap in it.

Miles: And Nelson's the same lesson.

Nora: Nearly two hundred years earlier. There was courage on both sides of that water, so courage alone can't explain it. What decided it was a read of where the other side was weak, on that day, in that sea, and a plan built entirely around it.

Miles: Who's Rumelt?

Nora: Richard Rumelt, born nineteen forty two. A bachelor's and a master's in electrical engineering at Berkeley, and from sixty three to sixty five a systems engineer at the Jet Propulsion Laboratory. Then a doctorate from Harvard Business School in nineteen seventy two, a spell on its faculty, and UCLA's management school from nineteen seventy six, where he's now emeritus. An engineer first, and it shows in a book that keeps asking what the mechanism is.

Miles: And the research his name rests on.

Nora: One paper above all. Nineteen ninety one, Strategic Management Journal, How Much Does Industry Matter. He took federal line of business reports on American manufacturing units for nineteen seventy four to seventy seven and split the differences in their profit rates into parts: the industry, the year, the parent company, and the business unit itself. The business unit won by a distance. A later review that tabulated his estimates puts the unit's own stable share at about forty four percent of the variance, and stable industry effects at about four.

Miles: So which business you are beats which industry you're in.

Nora: On that data, by roughly ten to one. Hold onto it: it's the closest the author comes to evidence that choices inside a firm decide things, and a map of where the differences sit rather than a test of what caused them. This book came out in July of twenty eleven, was a finalist for the Financial Times business book of the year, and a second book followed in twenty twenty two.

Miles: The argument, in his order.

Nora: Part one opens with what we've just told: good strategy is unexpected, because it's rare, and having one is an advantage before you've done anything with it. Chapter two is power: find where your strength meets somebody's weakness, and aim there.

Miles: Then the bad kind.

Nora: Four hallmarks. Fluff first: abstraction posing as insight. A retail bank's strategy, in its own memo, was customer centric intermediation. He translates it for you: a bank takes in deposits and lends the money out.

Miles: So the strategy is to be a bank.

Nora: Second, failing to face the challenge. A farm equipment maker's plan for nineteen seventy nine counted on market share and cost cuts, and never addressed what he calls the worst labour relations in American industry, the problem that mattered most.

Nora: Third, mistaking goals for strategy. A chief executive brings him a twenty twenty plan, twenty percent growth a year at twenty percent margins. Rumelt asks where the leverage will come from and is told the secret of success is setting your sights high.

Miles: Which is a target with a pep talk.

Nora: And fourth, bad targets: either a dog's dinner, his phrase, of things to do, or a blue sky line that restates the wish. His dog's dinner is a city in the Pacific Northwest whose plan held forty seven strategies and a hundred and seventy eight action items.

Miles: That's thorough.

Nora: Item a hundred and twenty two was to create a strategic plan.

Miles: A hundred and twenty one items in.

Nora: And still waiting for one. Then chapter four asks why there's so much of this, and gives three reasons. Strategy means choosing, and choosing disappoints somebody. His case is Digital Equipment in nineteen ninety two, three camps: keep selling boxes, sell solutions, or bet on chips. Nobody would pick, so the sentence everyone agreed to was a commitment to high quality products and services and to leading in data processing.

Miles: Nobody could vote against that.

Nora: Because it cost nobody anything. The second reason is template planning: fill in a vision, a mission, some values, then goals with the word strategies written over them. The third he calls New Thought, a line he traces from Emerson to the modern self help shelf, the belief that dwelling on obstacles invites them. If naming the problem counts as bad luck, nobody writes the diagnosis.

Miles: Does anything stand behind that third one?

Nora: Something stands next to it. Oettingen and Mayer, twenty oh two, Journal of Personality and Social Psychology, four studies that followed people forward in time. In the first, eighty three graduates hunting for work, forty of them reached two years later. Those whose daydreams of the search were rosiest had fewer job offers and lower pay. Judging success as likely, which is a different thing from enjoying the picture, went the other way. The same split turned up with students facing an exam, students with a crush, and patients after a hip replacement. The samples are small, and it's about daydreams rather than plans, so it supports his instinct without testing his claim.

Miles: And goals themselves? He sounds against them.

Nora: Less than a fast read suggests, and the goal research agrees with the narrow version. Locke and Latham, American Psychologist, twenty oh two, summing up thirty five years of work: specific hard goals beat do your best across more than a hundred tasks and over forty thousand participants. The same review says that on complex tasks, where nobody yet knows how, a hard performance goal can send people scrambling for methods at random, and a learning goal, such as finding a set number of different approaches, works better.

Miles: So the goal people back him.

Nora: On his exact point. A target is a weak instrument when the how is missing, and in strategy the how is usually what's missing.

Miles: Then the good kind.

Nora: Chapter five, the kernel, three parts. A diagnosis, which says what's going on and picks out the few things that matter. A guiding policy, the overall approach to the obstacle, which he likens to a signpost: direction without the details. And coherent actions, steps that support each other rather than pulling apart.

Miles: Give me one small enough to carry.

Nora: His friend Stephanie, who owns a corner grocery. She tinkers her way to a guiding policy, serve the busy professional, and then tightens it: the busy professional with little time to cook. Once that sentence exists, the decisions after it stop being separate arguments.

Miles: Part two.

Nora: Sources of power, nine chapters, and two I'd keep. The first comes from the lab he once worked at. The engineers designing a craft to soft land on the moon couldn't start, because nobody knew what the surface was like. A scientist there, Phyllis Buwalda, wrote them one: hard and grainy, slopes under fifteen degrees, scattered rocks no bigger than about two feet. He's clear that it wasn't the truth. The truth was that nobody knew. It was a target close enough to build to, the kind he calls proximate.

Miles: And if the moon disagreed.

Nora: Her answer, as he tells it, was that a much rougher moon wouldn't be one anybody spent long on anyway. The second is the chain link system, where the weakest part sets what the whole can do. He opens it with the seal that failed on the Challenger. Improve any other link and nothing moves.

Miles: Which is the constraint from five books back.

Nora: A close cousin. Then part three, thinking like a strategist. He calls a strategy a hypothesis, a judgment you test by acting, and reads Howard Schultz's espresso bars as an anomaly noticed, then tried. He offers mind tools: write the kernel down, build a plan and then try to wreck it, which he calls create and destroy, and picture a panel of experts judging it. And the last chapter uses the inside view, a term from the work of Kahneman, who arrives later in this run, to explain how herding kept people from seeing the two thousand eight crisis coming.

Miles: Now the evidence check.

Nora: Start with what kind of evidence this is, because the book doesn't hide it. It's cases, chosen from history and from his own consulting to show a pattern. That's illustration. A test would need the firms that had a sharp diagnosis and failed anyway, and the ones running on fluff that did fine, and the book doesn't go looking for either.

Miles: Which isn't the same as wrong.

Nora: No. It's the distance between a good argument and a measured one. And hindsight only runs one way. We know what Apple became after ninety seven, so waiting for the next big thing reads as patience. If the next big thing had never come, the same sentence would read as drift. A book near the end of this series is built entirely on that problem, so I'll leave the rest for it.

Miles: Has anybody measured anything close?

Nora: The nearest thing is strategic planning, which isn't quite his subject. George, Walker and Monster, Public Administration Review, twenty nineteen. A meta-analysis of thirty one studies and eighty seven correlations, public and private organizations together. Planning went with better performance, an average correlation of about point two three, which they call moderate. About three quarters of those studies were single cross-sectional surveys, one questionnaire supplying both the planning and the performance. The ones that drew on more than one source of data found smaller effects. That's last episode's problem turning up again in a new field.

Miles: So planning helps a bit, probably.

Nora: Planning goes with doing well, modestly. And look at what got measured: formal planning, close kin to the template he spends chapter four taking apart. Nobody has measured the kernel itself: whether organizations with a written diagnosis, a guiding policy and coherent actions outperform the rest. I went looking and found nothing that tests it.

Miles: And his own paper?

Nora: Supports his starting point and stops there. Differences between business units dwarf differences between industries, which leaves room for choices to matter. It can't show the choices made the difference, because luck, timing and everything nobody chose sit inside that same share.

Miles: Anything in the book you'd mark down?

Nora: One small thing that shows the genre. In chapter seventeen, Frederick Taylor tells Andrew Carnegie, in Pittsburgh in eighteen ninety, to list his ten most important tasks and start on the first, and a cheque for ten thousand dollars follows. The same story circulates about a different adviser, Ivy Lee, a different steel man, Charles Schwab, six tasks and twenty five thousand dollars. One writer who chased that version found nothing earlier than the nineteen sixties.

Miles: So it's a parable.

Nora: File it as one. The advice is good with or without a receipt.

Miles: The tier statement, then.

Nora: It's practitioner literature, not peer reviewed research, written by somebody whose peer reviewed research is real and sits underneath one corner of it. The kernel is a framework built from cases and decades of teaching and consulting, and it has never been put to a test. The planning evidence sits beside it, measures something else, and says mostly that planning goes modestly with doing well. His own word for a strategy is a hypothesis, and by that standard his cases are observations rather than results. Persuasive, useful and untested, and none of that's a verdict on the author.

Miles: Our half.

Nora: And it lives in what bad strategy asks of its readers. Fluff, a goal with no route, forty seven priorities: every one of them hands the reader the job of working out what the leaders meant. The people who do that job easily are the ones who already think like the leaders. Everyone else is guessing.

Miles: Is there evidence it's harder across that line?

Nora: There's a study I keep returning to. Crompton and colleagues, twenty twenty, in the journal Autism. Seventy two adults in nine chains of eight, a story passed from person to person. Chains of autistic people kept as much of the detail as chains of non-autistic people. The mixed chains lost noticeably more, and people in them reported lower rapport.

Miles: So the trouble is in the crossing.

Nora: Yeah. In the crossing, on that task. A small study, a story in a lab rather than a plan at work, and a group finding. It fits a pattern, though. Meaning that rides on shared intuition travels worse between people whose intuitions differ, and a written diagnosis doesn't ride on intuition at all.

Miles: And the price of not knowing.

Nora: Jenkinson, Milne and Thompson, twenty twenty, same journal. A meta-analysis of ten studies, five hundred and sixty two participants, almost all children and young people aged four to twenty four, questionnaires taken once. Intolerance of uncertainty and anxiety went together strongly, a correlation of about point six two, and both ran high in the autistic samples. The authors put the strength of that link at about what's been found in non-autistic samples. Not knowing what's expected costs people something, and on these measures many autistic young people carry more of that cost. A strategy that won't say what the problem is produces exactly that kind of not knowing, for everybody, all year.

Miles: And where does ADHD come in?

Nora: Through the guiding policy. Two books ago we called a checklist an external working memory. A guiding policy does that for decisions. Instead of holding every priority in your head and weighing each new request against all of them, you hold one sentence. Serve the busy professional with no time to cook. Does this help with that, yes or no.

Miles: So it suits anybody who can't run on vibes.

Nora: And it costs the people who can nothing. Buwalda's specification is the cleanest version in the book, and nobody would call it an adjustment. The engineers couldn't start without one, so somebody wrote it down, called it a best guess, and the work began.

Miles: Cautions.

Nora: Two. The first I owe every time. Nobody has studied strategy documents with neurodivergent staff as the outcome. I've put a communication study and an uncertainty review next to a management book and drawn the line between them myself. That's an argument. Call it one, and don't let it pass for a finding.

Miles: And the second.

Nora: Is about the word. Rumelt's diagnosis is of a situation: a market, a rival, a bottleneck, a cash position. It's never of a person, and on this show it never will be. The demand is that leaders write down what's wrong with the situation the plan faces. It isn't permission to start writing that down about the staff.

Miles: Fair enough.

Miles: A minute, in the terms this show uses. Good Strategy Bad Strategy says most of what passes for strategy is fluff, goals or long lists, and the real thing has a kernel: a diagnosis of what's in the way, a guiding policy for dealing with it, and actions that pull together. For anybody leading a mind that takes another route, or working inside one, here's Monday. Write the diagnosis down in two or three sentences before anybody mentions a target, so nobody on the team has to infer it. Turn the guiding policy into one sentence people can test a request against, since one sentence is cheaper to hold than forty seven priorities. Where you can't know, do what Buwalda did: write the assumption down, label it an assumption, and let the work begin. And when somebody asks what you mean, treat it as the strategy being tested, not a person being slow.

Miles: Now the tools.

Nora: The general caution about this class of tool has its own place after the music, in wording that never changes, so I'll leave it there and spend these minutes on this book.

Miles: Go on.

Nora: A checker, first. Paste in your strategy document and ask for his four hallmarks, sentence by sentence: which lines are fluff, which are goals wearing the word strategy, which list items have no route, and whether the document contains a diagnosis at all. It's a reading job, and you can check every flag yourself.

Miles: And past checking.

Nora: Create and destroy, his own mind tool. You draft the plan, then ask for the strongest case against it, or a panel of critics arguing from different seats. These tools are good at attack. Or rival diagnoses: give it the facts, ask for three different accounts of what's wrong, and for each one, what evidence would tell them apart. Then you choose.

Miles: Why not let it choose?

Nora: Two studies. The first you've met: the consultants' experiment three books back, where the tool lifted people on eighteen tasks and dragged them down on one. What we didn't spend time on then is what that one task was. Dell'Acqua and colleagues built it so that the right answer meant finding the root cause of a company's trouble from its performance data and interviews with its executives.

Miles: So the task it failed was a diagnosis.

Nora: Yeah. A diagnosis, the one part Rumelt says everything rests on, and it's the one that went backwards. The second is Doshi and Hauser, twenty twenty four, Science Advances. Two hundred and ninety three writers, randomly assigned to write a very short story alone or with ideas from one of these models, judged by six hundred readers. The assisted stories were rated more creative, and they came out more alike. It's fiction, not strategy, but his first chapter says good strategy works partly by being unexpected. If every rival asks the same kind of tool, the answers drift toward each other.

Miles: So the edge is in the part it can't do for you.

Nora: The diagnosis. A machine will write you a fluent guiding policy in seconds, and fluency is the danger, because fluff was always fluent. Let it check the document and wreck the plan. Saying what's wrong, in your market, this year, stays with you.

Miles: That's Good Strategy Bad Strategy.

Miles: Next time: The Innovator's Dilemma, by Clayton Christensen. Rumelt explains the choice, and Christensen explains why good firms, well run, still can't make it.

Nora: And it arrives with a public argument about its case studies, which I'll report as an argument, both sides.

Miles: That was a good one to start on, Nora. Thanks.

Nora: Glad we did that one.

Miles: Until the next one.

Miles: Sixty Books is put together at Stoar Consulting and Coaching. Logan Williamson is a Licensed Professional Counselor in Texas and Louisiana. You heard a single book read with care, for whoever leads a team and whoever works on one. It wasn't coaching, it wasn't counseling, and none of it was a judgment about you.

Nora: You'll find the source behind each claim on the episode page, with the full transcript. The next book arrives in two days.

Miles: For the record, twice over. These are synthetic voices. Miles isn't a person and neither is Nora, neither holds a licence of any kind, and nothing in this episode was said about whoever is listening. The scripts begin as drafts written with AI assistance, and none goes out until Logan Williamson, a Licensed Professional Counselor in Texas and Louisiana, at Stoar Consulting and Coaching, has edited it and given it his approval.

Nora: And one note about the AI segment. The evidence that these tools help neurodivergent people at work is early. There are controlled trials now, but they measure things like conversation skill rather than how a job actually goes, and for work itself most of what exists is individual providers reporting success. So take it as promising and not proven. The tools also get rebuilt every few months, so whatever you set up today is built on something that moves.