You said it clearly. You're sure you did. And three weeks later the work comes back looking nothing like what you asked for — not wrong exactly, just different, shaped by three people's separate interpretations of a sentence you thought was unambiguous.
Nobody disobeyed you. That's the disorienting part. Everyone in the chain did roughly what they thought you wanted. The thing that failed wasn't anyone's competence or loyalty. It was the chain itself.
Robert Greene's fifth strategy is about that chain — how orders degrade as they travel, how groups make worse decisions than individuals, and why the most tempting modern solution (poll everyone, decide together, be democratic about it) is the one that reliably produces disaster.1
The core claim is blunt and unfashionable: divided leadership is the cause of the greatest military defeats in history.1
Greene stacks up the evidence deliberately. At Cannae in 216 B.C., the Romans outnumbered Hannibal two to one and were annihilated — and much of the blame sits on a command system that gave two tribunes shared leadership of one army. They disagreed about how to fight Hannibal, fought each other nearly as much as they fought him, and made a mess of the whole thing.1
Two thousand years later, Frederick the Great outlasted five great powers partly because he made decisions faster than alliance generals who had to consult each other on every move. In World War II, Marshall insisted on a single supreme commander for the Allied armies — without winning that fight first, Eisenhower could not have succeeded in Europe. In Vietnam, Vo Nguyen Giap's unified command faced an American strategy assembled by a crowd of politicians and generals.1
The mechanism underneath is what Greene calls Groupthink, and his description of it is unsentimental.
People in groups are political. They say and do things calculated to protect their image within the group rather than to see the situation clearly. They aim to please, or to position themselves, rather than to be right. Where an individual can be bold, a group is usually afraid of risk. And the need to find compromise between competing egos kills creativity directly.1
The group has a mind of its own, and that mind is cautious, slow to decide, unimaginative, and sometimes downright irrational.1
Greene's prescription is not "be a dictator." It's more uncomfortable than that, because it involves a deliberate gap between what you do and what it looks like you're doing.
Preserve unity of command absolutely — the overarching strategic vision comes from you and you alone. Then hide your tracks. Work behind the scenes. Make the group feel genuinely involved: seek their advice, incorporate their good ideas, politely deflect the bad ones. Make minor cosmetic changes to the strategy where it soothes the insecure political animals in the room. But ultimately trust your own vision, and never actually relinquish the decision.1
This is the strategy's real ethical weight, and Greene states it without flinching. He is describing performed consultation — the appearance of democratic process laid over a decision that was never actually shared.
Control, Greene argues, is elusive in a specific way: the harder you pull at people, the less of it you have.1
Ingmar Bergman is the case. Early in his career he was overwhelmed by the gap between the films he could see in his head and the work of actually directing — so he lashed out, shouted orders, attacked his cast and crew for failing to deliver what he wanted. Some stewed with resentment. Others became obedient automatons. With nearly every new film he had to start over with a new crew, which made everything worse.1
What eventually fixed it wasn't better technique. It was assembling a permanent team — Sweden's finest cinematographers, editors, art directors, actors — who shared his standards and whom he trusted. With an actor like Max von Sydow, Bergman could suggest what he had in mind and watch it come to life. Greater control came from letting go.1
Greene is specific about the selection criteria, and the advice cuts against most hiring instinct: don't be seduced by expertise and intelligence.
Character matters more. The ability to work under you and alongside the rest of the team. The capacity to accept responsibility and think independently. That is why Marshall tested Eisenhower for months before giving him a command.1 Look past the résumé to the psychological makeup underneath it — a brilliant person who cannot operate inside your chain is worth less to you than a competent one who can.
Lincoln is the other case. He had a strategy for winning the Civil War and no military background to execute it, plus generals who disdained him. The strategy was worthless until he found Grant — a man who shared his belief in offensive warfare and, crucially, did not have an oversize ego. Once Lincoln found him, he latched on, gave him command, and let him run the war.1
And then the counterweight: rely on your team, but never become its prisoner. Roosevelt had his famous brain trust and depended on them for ideas and opinions — but he never let them into the actual decision making, and he deliberately prevented them from building independent power bases inside his administration. He saw them as tools that extended his own abilities. He understood unity of command and was never seduced into violating it.1
The single greatest risk to a chain of command, in Greene's account, comes from the political animals in the group. They are inescapable — they spring up like weeds in any organization.1
The damage they do is specific. They interpret your commands for their own purposes. They find loopholes in any ambiguity. They build factions to advance their agendas. And the breaks they create in the chain are invisible until something fails.
Greene's screening advice is practical. Look at candidates' histories — are they restless, do they move from place to place often? That kind of ambition tends not to fit in anywhere for long. And be wary of people who seem to share your ideas exactly: they are probably mirroring them to charm you.1
Elizabeth I's solution to a court full of political types was to keep her own opinions quiet. On any issue, nobody outside her inner circle knew where she stood, which made her impossible to mirror — you cannot disguise your intentions behind perfect agreement with a position you can't identify.1
The chapter closes on something easy to skip: orders have a form as well as a content, and the form does real work.
Vague orders are worthless. They get hopelessly altered as they pass person to person, and — worse — your staff reads them as symptoms of your own uncertainty. But over-specific orders are their own failure: they encourage people to behave like automatons and stop thinking, which is exactly what you need them to do when the situation changes.1
Napoleon is Greene's model here. His orders were full of concrete detail, which gave officers a feel for how his mind worked while still leaving interpretive room. He often spelled out contingencies, suggesting how an officer might adapt. And he made the orders inspiring — the language communicated the spirit of what he wanted, so that the recipient felt like a participant in something rather than a minion executing a distant emperor's wishes.1
You're two levels up from the people doing the work, and a project you care about is drifting. The instinct is to call an all-hands, lay out the problem, and get everyone's input on the fix.
Run Greene's diagnostic before you schedule it. Is this a genuine information-gathering problem — you don't know something the team knows — or is it a decision you're avoiding making? Those need opposite responses. The first genuinely benefits from the room. The second is where the meeting becomes a way of distributing responsibility for a call you should be making alone, and the output will be the cautious, compromised, everyone-can-live-with-it plan Greene is warning about.
If it's the second, do the harder thing. Decide. Then bring it to the room as a direction rather than a question — and mean the parts where you ask for input, incorporating what's genuinely good rather than performing consultation you've already discounted.
Then check the chain itself. Say the decision out loud to one person two levels down and ask them to repeat back what they'd do Monday. If what comes back isn't what you meant, the problem was never the team's execution. It was an order shaped like Hamilton's at Gallipoli — polite, general, and quietly disastrous.
The military evidence is genuinely strong and multi-sourced: Cannae, Frederick the Great, the Allied command structure in 1944, and Vietnam are four independent cases spanning two thousand years, and the pattern (divided command loses to unified command) holds across all of them.1 [POPULAR SOURCE] for Greene's framing.
The tension the chapter creates and doesn't resolve: Greene explicitly instructs you to manufacture the appearance of participation while retaining the reality of sole decision-making — "make the group feel involved in your decisions."1 He presents this as sophistication. It is also, described plainly, a recommendation to deceive your own team about how much their input actually matters.
The chapter never engages with the obvious cost: people generally find out. A team that discovers its consultation was theater doesn't just lose trust in the process — it loses the willingness to offer real information next time, which degrades exactly the intelligence flow the rest of the chapter says is essential. Greene's own Marshall case suggests a better answer he never articulates: Marshall didn't fake consultation, he actually delegated to people he'd trained to think like him, which produces the same speed without the deception.
Set this beside Strategy 6: The Controlled-Chaos Strategy, the very next chapter. Strategy 5 says: never relinquish unity of command, keep every string in your own hands. Strategy 6 says: break your forces into independent units, give them a mission, and let them run without waiting for your orders.
Read carelessly these contradict. Read carefully they don't, and the distinction is the useful part: Strategy 5 governs who sets the objective (one person, always), while Strategy 6 governs who decides how to reach it (whoever is closest to the situation). Greene never states this reconciliation directly, which leaves two adjacent chapters looking like opposite advice — and a reader applying either one without the other gets a predictable failure, either a bottlenecked micromanager or an organization with no coherent direction.
Psychology → Groupthink as Documented Phenomenon. Irving Janis's research on group decision-making identified the specific mechanism Greene describes — cohesive groups under pressure suppress dissent, overestimate their own judgment, and converge on worse decisions than their individual members would reach alone, with the Bay of Pigs as the canonical case.
The insight the pairing produces: Greene's prescription (unity of command) is one solution, but the research literature points at a different one he never considers — structured dissent, where someone is formally assigned to argue against the emerging consensus. Both fix the same problem. Greene's fix concentrates authority; the alternative distributes it while protecting against convergence. The chapter's confidence that concentration is the only answer is a claim it makes without ever testing it.
Business → Skip-Level Meetings and the Filtered Report. Every executive above a certain organizational size faces Hamilton's problem: information reaching them has passed through people with incentives to shape it, and the telling details get sanded off in transit.
The insight: what companies call "skip-levels" and Greene calls the directed telescope are the same instrument, and both exist because a chain of command is simultaneously the thing that makes an organization work and the thing that makes its leader progressively blinder the higher they climb. See The Directed Telescope for the dedicated treatment.
Everything starts from the top, and that includes the failures you'd most like to blame on someone else. If your orders arrive at the front as vague suggestions, the people who executed them vaguely are not the problem — you are, and the specific thing you did wrong happened weeks before anyone reached a beach.