Mission command
Leading by intent instead of by instruction: the Prussian origins (Auftragstaktik), the US Army's seven principles, how to write commander's intent, the five-paragraph order, troop leading procedures, backbriefs and rehearsals, after-action reviews, and how to run a startup this way without losing control of the things that must stay centralized. The decision cycle this speeds up is the OODA loop; the ownership mindset it needs is Extreme Ownership; delegation and one-on-ones in everyday management are in management.
The idea in one paragraph
Plans break on contact with reality, and the people closest to the problem see the change first. Mission command gives them the purpose and the desired result, the resources and the limits, and leaves the how to them, so that when the plan stops fitting they can act without waiting for permission. It only works with competent people, mutual trust and shared understanding, which is why it is as much a culture as a technique.
| Detailed command | Mission command | |
|---|---|---|
| Assumes | the world is predictable enough to plan in detail | the world is uncertain; plans will break |
| Orders say | what to do, step by step | what to achieve and why; minimum necessary constraints |
| Decisions made | at the top | at the lowest level with the information |
| Information flows | up for approval, down as instructions | intent down, situation up, sideways between peers |
| When the plan fails | stop, report, wait for new orders | act within intent, report when able |
| Strength | synchronization, predictability, control | speed, adaptability, initiative |
| Weakness | slow, brittle, overloads the top | needs trust and training; can desynchronise |
| Business analogue | approval chains, detailed specs, micromanagement | goals with guard-rails, empowered teams |
Origins: Auftragstaktik
| When | What happened |
|---|---|
| 1806 | Napoleon destroys the Prussian army at Jena–Auerstedt |
| 1807–1810s | reformers (Scharnhorst, Gneisenau, later Clausewitz) build a professional officer corps and a General Staff that educates officers to think, not just obey |
| 1857–1888 | Helmuth von Moltke the Elder is Chief of the Prussian (later German) General Staff; wins the wars of 1864, 1866 and 1870–71 with short directives and wide latitude for subordinates |
| 1871 | Moltke's essay Über Strategie (On Strategy): "No plan of operations extends with any certainty beyond the first contact with the main hostile force" |
| 1890s–1914 | German writers debate "free method" (freies Verfahren) and "mission method" (Auftragsverfahren); the word Auftragstaktik emerges from this debate |
| 1987 | British Army announces it will adopt mission command; an internal review in 2004 found its orders had become more detailed instead |
| 2003 | US Army FM 6-0 codifies the term "mission command" |
| 2012 → 2019 | ADP 6-0 renames the whole command function "mission command" (2012), then restores "command and control" and defines mission command as the Army's approach to it (2019) |
| July 2026 | current ADP 6-0 Command and Control supersedes the 2019 edition; mission command and its seven principles are unchanged; adds dispersed command posts, data and multidomain operations |
| today | the Bundeswehr calls it Führen mit Auftrag (leading by mission) |
The Moltke line in German reads "Kein Operationsplan reicht mit einiger Sicherheit über das erste Zusammentreffen mit der feindlichen Hauptmacht hinaus" (Hughes and Bell translation, Moltke on the Art of War, 1993). "No plan survives contact with the enemy" is a later paraphrase, not his wording. His point is narrower and more useful: plan in detail up to the first decisive clash, and beyond that plan only the intent and the options.
ADP 6-0's favorite American example is Grant's letter to Sherman of 4 April 1864: "I do not propose to lay down for you a plan of campaign, but simply to lay down the work it is desirable to have done and leave you free to execute it in your own way. Submit to me, however, as early as you can, your plan of operation." That is intent, freedom of action and a backbrief in three sentences. Sherman's reply of 10 April restated the aim (knock Johnston's army out and damage the enemy's war resources, keeping Johnston too busy to reinforce elsewhere) and promised not to "let side issues draw me off from your main plan".
The seven principles (ADP 6-0)
ADP 6-0 defines mission command as "the Army's approach to command and control that empowers subordinate decision making and decentralized execution appropriate to the situation". Its principles, identical in the 2019 and 2026 editions:
| Principle | What doctrine means | Business translation | Warning sign it is missing |
|---|---|---|---|
| Competence | tactically and technically competent leaders and teams, built by training and education | people are trained and experienced enough to be trusted with the decision | "empowering" someone who has never done it and has no support |
| Mutual trust | trust up and down, earned through shared experience, two-way communication and keeping promises | the boss backs good-faith decisions; the team tells the boss the truth | people ask permission for things inside their authority |
| Shared understanding | a common picture of the situation, the problem and the approach, pushed to the lowest level | everyone sees the same metrics, customer feedback and strategy | teams surprised by decisions others consider obvious |
| Commander's intent | a clear, concise statement of purpose, key tasks and end state | the why and what-good-looks-like for a project or quarter | work that meets the spec and misses the point |
| Mission orders | directives that state the results to attain, not how to achieve them | goals and constraints, not task lists | 30-step plans; "just do what the ticket says" |
| Disciplined initiative | the duty to act within intent when orders no longer fit, the situation changes or an opportunity appears | people fix the problem and tell you, rather than wait | people follow a failing plan because it is the plan |
| Risk acceptance | commanders deliberately accept and underwrite risk, including subordinates' good-faith mistakes | leaders absorb the cost of reasonable bets that fail | the first honest mistake gets someone publicly punished |
Disciplined initiative: when to depart from the plan
ADP 6-0 tells subordinates to follow the plan until it no longer fits, then act, and to weigh two things:
- Do the benefits of the action outweigh the risk of desynchronising the overall operation?
- Will the action further the higher commander's intent?
If there is time, tell the boss the new situation and the recommended action first. If there is not, act within intent and report when able. Doctrine's tie-breaker: if in doubt whether to wait or seize a fleeting opportunity, act if you can do so within the intent.
Commander's intent
ADP 6-0 (quoting JP 3-0): "a clear and concise expression of the purpose of an operation and the desired objectives and military end state". It includes three parts, and must be understood by people two levels down.
| Part | Meaning | Test |
|---|---|---|
| Purpose | the broader why, in relation to the higher leader's intent (not a restatement of the mission's "why") | if the plan collapses, does this tell people which way to go? |
| Key tasks | the few things the whole force must do to reach the end state; not a task list for any one team | would dropping one of these make success unlikely? |
| End state | the conditions that exist when the operation is complete: ours, the competitor's, the customer's, the market's | could two people independently agree whether we got there? |
Doctrine's style rules: the leader writes it personally, delivers it face to face where possible, keeps it short enough to remember, and includes the purpose of each key task so people can improvise sensibly.
Business intent template
INTENT: <project / quarter / initiative>
PURPOSE Why this matters to the company's goal this year,
in one or two sentences. Link to the level above:
"So that <company goal>, we need to..."
KEY TASKS (3 max, each with its purpose)
1. <task> - so that <reason>
2. <task> - so that <reason>
3. <task> - so that <reason>
END STATE (observable conditions when we are done)
- Customers: ...
- Product/system: ...
- Team/company: ...
CONSTRAINTS Must do / must not do (budget, legal, brand,
security, dates that cannot move).
DECISIONS You decide: ... Check with me: ...
BACKBRIEF Tell me your plan by <date>.Good and bad intent
| Weak intent | Why it fails | Stronger intent |
|---|---|---|
| "Launch v2 by June." | a date and an output; no purpose or success condition | "Launch v2 so that mid-market customers can self-serve onboarding. Done when 70% of new mid-market accounts activate without a CSM call, with no increase in support tickets per account. Date matters less than quality: we'd rather slip two weeks than launch something that needs hand-holding." |
| "Grow revenue." | unbounded; every action qualifies | "Grow net new ARR 25% this quarter without discounting over 15%, so we reach the Series B milestone with healthy unit economics. Key tasks: fix the trial-to-paid drop at step 3; land 5 reference customers in healthcare." |
| "Improve reliability." | nobody can tell when it is done | "Get checkout availability to 99.95% for a full month so enterprise prospects stop raising reliability in procurement. We'll trade feature work for it, not headcount." |
Mission orders and levels of control
A mission order is not a separate kind of order but a way of writing one: state the mission, task organization, intent, concept of operations, tasks to subordinate units and essential coordinating instructions, with the emphasis on purpose. It never repeats what standard procedures already cover, and never tells a subordinate how to do their task. The right level of detail depends on the situation.
When tighter, centralized control is correct. ADP 6-0 lists operations needing close synchronization (combined arms breaches, air assaults, river crossings) as ones to control more tightly, and pursuits or movements to contact as ones to loosen. Factors that push toward more control: an enemy's capability, the degree of synchronization needed, constraints from above, risk, legal and ethical ambiguity, rules of engagement, and low cohesion, training, trust or shared understanding.
| Situation in a company | Control | Why |
|---|---|---|
| security breach, legal or regulatory exposure, press crisis | tight, one incident commander | errors are irreversible; one voice |
| database cutover, pricing change, coordinated launch | tight during the synchronized phase, loose before and after | many teams must move at the same moment |
| new team, new hires, or someone doing this for the first time | tighter, with coaching | competence and trust not yet built |
| a 3-person company | little formal structure needed | shared understanding is automatic |
| product discovery, sales prospecting, content, most engineering work | loose: intent, constraints and a check-in | local knowledge beats central planning |
| an experiment with a small, capped downside | loosest | the point is to learn fast |
Mission orders do not mean no supervision. Commanders remain accountable, check on subordinates and intervene when needed; in training, they deliberately intervene only when necessary so people learn.
The five-paragraph order
The Army standard is the five-paragraph operation order (OPORD): Situation, Mission, Execution, Sustainment, Command and signal. The Marine Corps version is taught as SMEAC, with the fourth paragraph "Administration and logistics".
| Para | Military content | Business version |
|---|---|---|
| 1. Situation | area of operations, enemy forces, friendly forces (higher unit's mission and intent, adjacent units), civil considerations, attachments, assumptions | market and customers; competitors; our position; what other teams are doing that affects us; assumptions |
| 2. Mission | who, what, when, where, why, in one sentence | "The growth team (who) increases trial-to-paid conversion from 8% to 11% (what) across all self-serve plans (where) by 30 September (when) so that we hit the ARR target without more sales hires (why)." |
| 3. Execution | commander's intent; concept of operations; tasks to subordinate units; coordinating instructions | intent (purpose, key tasks, end state); the plan in phases; who owns what; shared rules (dates, review points, decision thresholds) |
| 4. Sustainment / admin and logistics | supply, maintenance, medical, personnel | budget, tools, vendors, hiring, on-call, who covers absences |
| 5. Command and signal | where the commander is, succession of command, communications plan | owner and deputy; decision rights; meeting cadence; channels; escalation path; how progress is reported |
Business OPORD template
1. SITUATION
Customers/market: ...
Competitors ("enemy"): ...
Us ("friendly"): higher-level goal and intent: ...
adjacent teams and what they're doing: ...
Assumptions: ...
2. MISSION
<Who> <does what> <where> <by when> <so that why>.
3. EXECUTION
Intent: purpose / key tasks / end state
Concept: phase 1 ..., phase 2 ..., phase 3 ...
Tasks: Team A ... so that ...; Team B ... so that ...
Coordinating instructions: review dates, thresholds
("if conversion < 9% by week 6, we regroup"), rules.
4. SUSTAINMENT
Budget, tools, people, vendors, on-call.
5. COMMAND AND SIGNAL
Owner: ___ Deputy: ___ Decides alone: ___
Escalate: ___ Cadence: weekly 30 min, Tue
Channel: #proj-___ Status: Friday written updateA warning order (WARNORD) is the short heads-up issued early so people can start preparing before the full order; a fragmentary order (FRAGORD) changes an existing order without reissuing it. In a company: the Slack message on Monday ("we're going to do X, full plan Thursday, start thinking about Y") and the mid-project change note.
Troop leading procedures
FM 5-0 (2024) sets out eight troop leading procedures (TLP) for small units without a staff. Steps 1–3 are usually in order, 3–7 are sequenced as needed, and step 8 runs throughout.
| Step | Military | Condensed business planning |
|---|---|---|
| 1 | Receive the mission | get the goal; assess the time available |
| 2 | Issue a warning order | tell the team what's coming so they can start preparing |
| 3 | Make a tentative plan | analyze the problem; sketch a plan |
| 4 | Initiate movement | start the long-lead work now (hiring, vendor contracts, data access) |
| 5 | Conduct reconnaissance | go and look: talk to customers, read the code, check the data |
| 6 | Complete the plan | refine with what you learned; write it up |
| 7 | Issue the order | brief it face to face; take questions; confirmation briefs |
| 8 | Supervise and refine | backbriefs, rehearsals, check-ins, adjust |
The one-third, two-thirds rule (FM 5-0): use at most one-third of the time before execution for your own planning and leave two-thirds for your subordinates to plan and prepare theirs. In a company: if the quarter starts in 3 weeks, the leadership plan and intent should be out within a week, not on the last Friday.
Backbriefs, confirmation briefs and rehearsals
| Tool | What it is | Business version |
|---|---|---|
| Confirmation brief | immediately after receiving an order, the subordinate briefs back their understanding of the intent, their tasks and purposes, and how they relate to others | "Tell me in your own words what we're trying to achieve and what your team owns." (2 minutes, end of the kickoff) |
| Backbrief | later, the subordinate briefs how they intend to accomplish the mission; the leader checks it fits the intent and the plans of adjacent teams | each team lead presents their plan to the leader and the other leads a week after kickoff |
| Rehearsal | walking through the plan before execution; from a key-leader talk-through on a map to a full-dress run | a launch dry-run, a game day, an incident drill, a mock customer demo |
Bungay's three gaps
Stephen Bungay, The Art of Action (2011), reframes mission command for business around three gaps between plans, actions and outcomes, and argues that the instinctive fix for each makes it worse:
| Gap | Between | Instinctive (wrong) response | Mission command response |
|---|---|---|---|
| Knowledge gap | what we'd like to know and what we actually know | gather more information, plan in more detail | don't plan beyond what circumstances allow; define and communicate the intent |
| Alignment gap | what we want people to do and what they actually do | give more detailed instructions | let each level work out how to achieve the intent and backbrief it |
| Effects gap | what we expect our actions to achieve and what they actually achieve | add more controls and metrics | give people freedom to adjust their actions in line with the intent |
He calls the resulting approach directed opportunism: strategy sets the direction; people seize opportunities within it.
Backbrief agenda (45 minutes, one team)
1. Restate the higher intent in your words. (5)
2. Your mission and intent for your team. (5)
3. Your plan: phases, owners, key dates. (10)
4. Dependencies on other teams; what you need. (5)
5. Assumptions and the biggest risks. (5)
6. Decision points: what triggers a change. (5)
7. Leader and peers: questions, conflicts, gaps. (10)
Output: agreed changes to the plan or the intent.After-action reviews
The US Army's current guide is TC 7-0.1 After Action Reviews (February 2025); the older TC 25-20 (1993) popularised the familiar four-question form. An AAR is "a guided analysis of an organization's performance, conducted at appropriate times during and at the conclusion of a training event or operation, with the objective of improving future performance". Two rules from TC 7-0.1: an AAR is not a critique, and it does not grade success or failure.
| Question | TC 7-0.1 section | What happens |
|---|---|---|
| What was supposed to happen? | commander's mission and intent | the leader (or subordinates) restate the plan and intent |
| What actually happened? | summary of key events | build one agreed timeline from many perspectives and objective data before any analysis |
| Why did it happen? | root causes | open-ended questions; causes emerge from the discussion, not from the facilitator |
| What do we sustain and improve? | sustaining / improving performance | specific actions, owners and dates; update procedures |
Facilitation guide
Before
- Hold it during or immediately after the event; the shorter the gap, the better the recall.
- Pick 3–5 focus topics in advance; gather objective data (timeline, dashboards, logs, recordings).
- Choose a facilitator who was not the decision-maker, if possible.
Open (5 minutes)
State the rules: participation expected; not a critique; no grading; rank does not win arguments; we're here to improve, not to assign blame. The senior person goes first with something they would do differently.
Walk the four questions (30–45 minutes)
Establish what was supposed to happen, then build the agreed timeline, then ask why. Ask open questions of the most junior people first: "What did you see when the alert fired?" beats "Why didn't you roll back?". Keep a visible list of causes and of things that went well.
Close (10 minutes)
Agree actions with an owner and a date each; record them; decide what changes in the standard procedure or template. Summarize in writing within 24 hours.
Follow up
Track actions to completion in the normal work system; review the list at the next AAR. An AAR whose actions never get done teaches people that AARs are theatre.
AAR record
EVENT: ______________ DATE: ______ FACILITATOR: ______
1. Intended: mission, intent, success criteria
2. Actual: timeline (time - event - source)
3. Why: causes (systems, information, decisions, luck)
Went well and why:
4. Sustain: ...
Improve: action - owner - due date
5. Changes to procedures/templates: ...Blameless postmortems
Google's SRE book (chapter 15, "Postmortem Culture: Learning from Failure") is the software form of the AAR. Postmortems are triggered by criteria agreed in advance (user-visible downtime beyond a threshold, any data loss, on-call intervention such as a rollback, resolution time beyond a threshold, a monitoring failure), and they must be blameless: focused on contributing causes, assuming everyone acted with good intentions on the information they had. If people are punished for what a postmortem reveals, they stop revealing things.
Pre-mortems and red teaming
| Tool | What it is | Use before |
|---|---|---|
| Pre-mortem | Gary Klein (Harvard Business Review, 2007): imagine the project has already failed and ask what went wrong | committing to any big plan; step 6 of TLP |
| Red team | an independent group that challenges plans, assumptions and the leader's picture, often by thinking as the adversary | strategy decisions, security, major bets |
The US Army formalised red teaming after a 2003 Defense Science Board report: it stood up a service-level red team in 2004, and the University of Foreign Military and Cultural Studies (UFMCS, the "Red Team University") at Fort Leavenworth trains red-team leaders and members. For a startup, the cheap version is a named devil's advocate for each major decision and an outsider (advisor, friendly customer) asked to attack the plan. How to run both: decision-making.
Two levels up, one level down
The doctrine behind the slogan:
- Intent nests. Each leader's intent nests within the intent one and two echelons above (FM 5-0), and understanding intent two echelons up is what lets people act sensibly when their own boss is unreachable (ADP 6-0).
- Intent carries down. A commander's intent must be easy to remember and clearly understood two echelons below (ADP 6-0).
The common rule of thumb that follows: know your boss's boss's goal, and know your team's work in detail.
| Level | You should know | Startup example for an engineering lead |
|---|---|---|
| two up | the company's goal and why | "reach $5M ARR with healthy margins before the next raise" |
| one up | your boss's intent and constraints | VP Eng: "cut infrastructure cost 30% without hurting reliability" |
| you | your mission and intent | "move batch jobs to spot instances by Q3" |
| one down | each team member's tasks, blockers and skills | who owns the scheduler rewrite, who needs help with Terraform |
| two down | enough to check the intent survived translation | ask a junior engineer why the migration matters |
Span of control
ADP 6-0 defines span of control as the number of subordinates or activities under one commander, and gives no magic number: in fluid, fast-changing situations a commander can effectively supervise fewer subordinates; more subordinates add flexibility. Narrowing the span adds layers, and more layers slow information up and down; widening it flattens the organization but overloads the leader. Extreme Ownership puts the practical limit at around six to ten people. For startups, the usual failure is a founder with 15 direct reports who has become the bottleneck for every decision: add a layer or push decisions down with explicit intent.
Applying it to a startup
Writing intent for a quarter
Q3 INTENT (CEO, 20-person B2B SaaS)
PURPOSE Prove we can sell to mid-market without founder-led
sales, so the Series A story is "repeatable growth",
not "great founders".
KEY TASKS
1. Close 12 mid-market deals with no founder on the call
- so that we know the sales motion is transferable.
2. Cut onboarding to under 7 days - so that deals don't
stall between signature and value.
3. Keep gross churn under 2%/month - so that new growth
isn't hiding a leak.
END STATE Two AEs each closing 1+ deal a month without
founders; onboarding median under 7 days; a
documented playbook a new AE can follow.
CONSTRAINTS No discounts over 20% without CEO sign-off.
Security review answers must come from eng.
DECISIONS Team leads decide their own plans, hires within
budget, tooling under $500/month.
BACKBRIEF Each lead presents their plan by 10 July.Delegation levels
Jurgen Appelo's seven delegation levels (from Management 3.0, 2011) make "decide within intent" explicit. Agree the level per decision type, not per person.
| Level | Name | Meaning |
|---|---|---|
| 1 | Tell | I decide and tell you |
| 2 | Sell | I decide and try to convince you |
| 3 | Consult | I ask for your input, then decide |
| 4 | Agree | we decide together |
| 5 | Advise | I give advice, you decide |
| 6 | Inquire | you decide, then tell me |
| 7 | Delegate | you decide; I don't need to know |
A typical early-stage map: pricing and hiring bar at 3–4; feature scope within the roadmap at 6; bug fixes, tooling and internal process at 7. More in management.
A mission-command operating rhythm
| Cadence | Mechanism | Mission command purpose |
|---|---|---|
| quarterly | written intent from the CEO; each lead writes nested intent; backbriefs | intent and alignment |
| weekly | 30-minute leads sync: what changed, what I decided, what I need | shared understanding; disciplined initiative reported up |
| daily | metrics everyone can see | shared understanding at the lowest level |
| after every launch, incident or lost deal | AAR or blameless postmortem | learning; trust |
| as needed | FRAGORD-style change note: what changed, new intent if any | keep the intent current without re-planning everything |
Failure modes
| Failure | What it looks like | Fix |
|---|---|---|
| Intent too vague | "make the product great"; people optimize for different things | purpose, 3 key tasks, an observable end state; test by asking two people if they'd agree it's done |
| Intent that is really a task list | ten bullet points of what to do | cut to the why and what-done-looks-like; move tasks into the teams' own plans |
| No shared context | intent is clear but people lack the information to apply it | open dashboards, customer calls, strategy docs; push information down, not just intent |
| Micromanagement dressed as intent | "you own it" followed by daily detailed direction and overrides | agree the decision level up front; intervene only on the pre-agreed triggers |
| Abdication dressed as empowerment | intent issued, then no backbrief, no check-in, no support | backbriefs, rehearsals and supervision are part of mission command, not a violation of it |
| Initiative punished | the first good-faith mistake is criticized in public | underwrite honest mistakes; review them in a blameless AAR |
| Competence assumed | juniors given senior-level latitude with no training | tighten control for new people and new problems; loosen as competence is proven |
| Desynchronisation | teams each act sensibly and collide (two launches the same day, conflicting promises to a customer) | coordinating instructions; peers backbrief together; tighter control during synchronized phases |
| Plans that are too detailed to adapt | the plan is followed after it stopped making sense | plan in detail only to the first contact; beyond that, intent and options |
| AARs as theatre | reviews held, actions never done | owners and dates; track in the normal work system; revisit next time |
References
- US Army, ADP 6-0 Command and Control (July 2026) (opens in a new tab): current doctrine; definition and seven principles of mission command, commander's intent, levels of control, span of control, Grant–Sherman vignette; supersedes ADP 6-0 Mission Command (31 July 2019)
- US Army, FM 5-0 Planning and Orders Production (November 2024, incl. change 1) (opens in a new tab): mission statement, commander's intent, five-paragraph order, troop leading procedures, one-third/two-thirds rule, confirmation briefs and rehearsals
- US Army, TC 7-0.1 After Action Reviews (February 2025) (opens in a new tab): AAR definition, rules, agenda and facilitation techniques
- US Army, FM 3-0 Operations (2025): the operational concept that mission command serves
- US Marine Corps, MCDP 1 Warfighting (1997): maneuver warfare, commander's intent and focus of effort in Marine doctrine
- Stephen Bungay, The Art of Action: How Leaders Close the Gaps between Plans, Actions and Results (Nicholas Brealey, 2011): the three gaps, directed opportunism, briefing and backbriefing
- Daniel J. Hughes (ed.), Moltke on the Art of War: Selected Writings, trans. Daniel J. Hughes and Harry Bell (Presidio, 1993): source translation of "On Strategy" (1871)
- Wikiquote: Helmuth von Moltke the Elder (opens in a new tab): the original German, the Hughes translation, and the later "no plan survives contact" paraphrases
- Wikipedia: Mission-type tactics (opens in a new tab): Prussian origins, the terminology debate and adoption by other armies
- Google, Site Reliability Engineering, chapter 15: Postmortem Culture: Learning from Failure (opens in a new tab): postmortem triggers and blameless culture
- Gary Klein, "Performing a Project Premortem", Harvard Business Review (September 2007) (opens in a new tab): the pre-mortem technique
- Wikipedia: Red team (opens in a new tab): the 2003 Defense Science Board recommendation, the Army's 2004 red team and UFMCS training
- Jurgen Appelo, Management 3.0: Leading Agile Developers, Developing Agile Leaders (Addison-Wesley, 2011): the seven levels of delegation
- Jocko Willink and Leif Babin, Extreme Ownership (St. Martin's Press, 2015): Decentralized Command and span of control from a small-unit perspective