Mission command

Why does an army built on obedience train its junior officers to depart from the plan?

An army spends a recruit's first weeks teaching him to do precisely what he is told, precisely when he is told it. Then it takes the same man, makes him a lieutenant, and tells him that when the plan stops fitting the ground he is expected to depart from it. That looks like an institution contradicting itself. Unless the obedience and the departure are answers to two different problems — in which case, what are they?

Start with where the information actually is. A commander two levels up has breadth: he knows the neighbouring units, the reserve, the timetable, the purpose. The officer at the front has something else — depth and currency. He can see that the ford is impassable, that the wood is held, that the enemy is already withdrawing. That knowledge is not in the plan because it did not exist when the plan was written.

So there are two ways to use it. Send it up, wait for a revised order, and act. Or act on it now, in a way the commander would have chosen had he known. Which is better depends on a comparison we can actually make: the cost of an error against the cost of a delay.

Follow the first route through. The report is composed, transmitted, queued behind other reports, read, judged, turned into an order, transmitted back, understood — and the situation does not wait. If the ground changes faster than that loop turns, the order that arrives answers a question nobody is asking any more. The delay does not merely slow you down; it degrades the decision, because it is made on a picture that has expired.

That gives the rule its shape: where the situation changes faster than the decision loop, delegate. And it tells you what to delegate. Not the purpose — the commander owns that, because only he can see how the parts fit. Only the method. So the order communicates intent: what is to be achieved, why it matters, what the acceptable risk is, and what must not happen. The method is left to whoever is standing there.

Now the obvious objection. If five officers each improvise, do they not improvise in five incompatible directions? They would, unless something makes their choices predictable to one another. That is what shared doctrine supplies: a common set of tactical habits, so that I can guess what my neighbour will do about a blocked road without asking him. Doctrine here is not the opposite of initiative — it is the thing that makes initiative safe to grant.

And there is a second requirement, which is harder. A subordinate acting on better local information will sometimes be wrong. If honest errors made in pursuit of the stated intent are punished, the rational subordinate stops deciding and starts waiting — which is exactly the behaviour the system was built to remove. So the arrangement needs trust running both ways, and trust is not a policy. It is a record of past behaviour.

Which explains something otherwise puzzling: why armies that admire this approach so often fail to adopt it. You can publish the doctrine tomorrow. You cannot publish the decade of shared training that makes improvisation mutually legible, or the accumulated evidence that initiative will be backed rather than charged. An organisation that has spent years rewarding compliance has taught its officers a lesson no memorandum undoes — and modern communications make it harder still, since a senior commander who can watch a live feed is under constant temptation to use it.

Mission command is not looser orders, and often not shorter ones. Stating a purpose, the constraints on it, and the risk you are willing to run takes more words than a route and a time, not fewer. What shrinks is the specification of method.

Nor is it licence to disregard orders one dislikes. The freedom is over means, bounded by the stated intent and by explicit limits. It bites in the specific case where the ordered method has stopped serving the stated purpose — and it can only bite if the purpose was actually stated, which is why doctrines of this kind insist that officers understand the intent two levels above their own.

Two honest qualifications. The historical story is tidier in the retelling than it was in fact: the practice is usually traced to nineteenth-century Prussian reform and to Moltke's habit of issuing directives rather than detailed orders, but the label Auftragstaktik came into common use later than the practice, and how consistently any army has ever managed it is genuinely argued among historians. And it is not universally correct. It trades control for tempo, so wherever a local optimisation can cause a global disaster — deconfliction of fires, air defence, anything touching escalation — tight central control is the right answer, and armies that practise mission command say so explicitly.

• How the same logic reappears in civilian organisations under the name of decentralisation, and why it fails there for the same two reasons.
• What happens to delegation when a commander can watch a live feed of the decision he delegated.

Every term the collection defines is gathered in the glossary.