Lexicon

Strategy and decisions

Strategy

Sir Lawrence Freedman, in Strategy: A History (2013), calls strategy "the art of creating power": getting more out of a situation than your starting position would suggest. Other people have their own aims and will act on them, so the situation keeps moving and the strategy has to move with it. Our preferred model is ends, ways and means, the formulation Arthur Lykke taught at the US Army War College from the late 1980s: define the objectives (ends), identify the methods to achieve them (ways) and allocate the resources (means). This provides a framework within which to make and maintain a plan. Without that framework, "priority" and "strategic" become labels for whatever feels important on any given day. A plan is one route from here to there; when it needs to be adjusted, the ends, ways and means provide a handrail against which to adapt.

Decisions and dilemmas

A decision is a choice between options that have real merit; the team picks the one that moves it closer to its ends. A dilemma is a choice between options that all carry a cost; we are forced to pick the one which costs the least. The distinction matters because "make better decisions" is the wrong instruction for a team facing a dilemma. Deciding in good time is how a team avoids being left with one.

Problem scoping

Problem scoping is the deliberate action of working out what exactly it is that requires a solution. It nests inside strategy, particularly when working to define appropriate ends. Others call it problem framing or problem definition. Framing for us is setting the boundaries in which we can think and act; definition is then understanding exactly what the problem is within the frame. In Cynthia Atman's studies of engineering design, experts spent significantly longer scoping a problem than students did, and gathered more information before they started to design. Most organisations do it less well: 85% of the executives Thomas Wedell-Wedellsborg surveyed said their organisations were bad at diagnosing problems.

Mission command and commander's intent

Mission command is the command philosophy in which a leader tells the team what they are trying to achieve and why (the mission), then trusts them to work out the how. Codified in the Prussian and German armies as Auftragstaktik, it is now the foundation of British Army doctrine. The commander's intent is the short statement of purpose that travels with the order and survives when the plan does not. The US Marine Corps describes it as what lets subordinates depart from the plan when the unforeseen happens and still serve their commander's aims. It is our favoured approach because it keeps a team effective when contact with reality breaks the plan. The distinction is between intent and instruction: an instruction tells people what to do; an intent tells them what the instruction was for.

The OODA loop

John Boyd's decision cycle: observe, orient, decide, act. The team that runs the cycle faster than its opponent keeps presenting the opponent with a situation that is already changing. What the popular version misses is that the critical step is orientation, not speed. A fast loop with poor orientation is a quick way to be wrong. Teams and leaders who take part in effective decision making exercises can invest in the orientation step.

Manoeuvre and attrition

Two ways to win a contest. Attrition is to be bigger and grind the other side down. Manoeuvre is to outthink and outpace them, to put them in positions where their plan no longer works. Some attritional contact is usually unavoidable; manoeuvre is about making sure you only fight where and when you hold the advantage. It depends on mission command: the openings appear and close faster than orders can travel, so a team has to understand the intent well enough to take them without waiting to be told. The modern Western articulation is William Lind's Maneuver Warfare Handbook (1985), through which Boyd's thinking reached the US Marine Corps and NATO. Outside the battlefield the distinction is between competing on resources and competing on tempo and position.

The 40 to 70 rule

Colin Powell, the US Army General turned Secretary of State, recommended action on somewhere between forty and seventy per cent of perfect information: below forty is a guess, and beyond seventy the situation will no longer be yours to decide. Nobody can measure what share of the information they hold, so the numbers are a guide to timing rather than a calculation. The rule gives permission to decide on an incomplete picture, and a handrail for when waiting for longer will not help. A competitor, an adversary or a market all move on their own clock, so the 70% solution in good time beats the 100% solution that arrives late. Wait too long and a decision becomes a dilemma.

Stakeholders

The four stakeholder groups

A stakeholder map drawn without a strategy is a cast of players. Drawn against clear ends, ways and means, it sorts the people on it into four groups: who can stop you, who you need, who could help if they chose, and who can only watch. The first two are the ones most maps find. The third is the one most maps miss, and it is where collaboration lives. The fourth is the one most maps waste effort worrying about. Whichever group a name sits in, their categorisation is just a hypothesis about another person until that person has been roleplayed in an exercise, giving them a voice and a decision to make.

Foresight

Horizon scanning

Horizon scanning is the discipline of keeping watch for change before it arrives: naming the forces moving around the organisation, saying which are moving fastest, and deciding what indicator would signal movement. The point is not to predict the future but to be ready to make effective, efficient decisions in good time.

Likelihood and impact

Aaron Wildavsky distinguished anticipation, which tries to predict specific harms and prevent them, from resilience, which builds the capacity to cope with harms as they arrive (Searching for Safety, 1988). A risk regime that ranks by likelihood is doing anticipation; its product is a thick plan against the expected list. An exercise rehearses the handling of consequences, deciding under pressure, communicating, re-planning, and that capacity carries across whichever risk fires. The distinction: likelihood tells you what to expect; impact tells you what you will have to do. We work in the impact space, building resilience against whatever comes out of our blind spot.

Resilience and crisis

Contingency planning

Contingency planning is a proactive process of thinking about and preparing for the moment that Plan A has to become Plan B. In its narrow sense, the one the Business Continuity Institute uses, a contingency plan is a set of pre-defined actions for a specific, named risk: the supplier fails, the site floods, the system goes down. Those plans have their place, alongside risk management and the business continuity plan. We use the broader sense: resilience comes from the teamwork, processes and agility to deal with whichever event arrives. A team that has practised deciding, communicating and re-planning under pressure is ready for the flood, the outage and the pandemic, because the skills that carry the response are the same in each.

Work-as-imagined and work-as-done

Work-as-imagined is how designers, managers and auditors assume the work happens, or should. Work-as-done is what the people at the sharp end actually do to get it done, with the time, kit and information they really have. The idea comes from French ergonomics in the 1950s, which separated prescribed work from real work; Erik Hollnagel brought it into resilience engineering, and Steven Shorrock has since added work-as-prescribed and work-as-disclosed (what people say they do). Hollnagel's point is that the adjustments people make to close the gap are why things usually go right, as well as why they sometimes go wrong. Exercising can bring that sharp-end expertise and practical rehearsal back to the plan, to stress-test it before it meets the practical constraints of execution.

Test to destruction

The purpose of stress-testing a plan in a wargame or exercise is to identify the point at which it breaks. We push the plan until it fails and learn from that failure. We can then either refine the plan to make it more resilient or keep watch for those factors we know we can't manage. A scenario the plan survives comfortably is just a demonstration and tells us little.

Red teaming and wargaming

Wargame

Peter Perla's definition of wargaming, from The Art of Wargaming (1990), is the one most practitioners use: a model or simulation of warfare, run without real forces, in which the flow of events shapes, and is shaped by, the decisions of players representing the opposing sides. Two things in it matter to us. The players are on opposing sides, so the opposition thinks and reacts instead of following a script. And the decisions drive the events, so the players see the consequences of what they chose and have to deal with them. We use the definition outside defence too, where the opposing side is a competitor, a regulator or a market. Without modelled opposition, what you have is a BOPSAT: a bunch of people sat round a table.

Red teaming

Red teaming, as the MOD's Red Teaming Handbook (DCDC, third edition, 2021) describes it, brings an independent view and a set of structured, creative and critical thinking techniques to a plan, so that whoever owns it decides with better information and ends up with something more robust. Search for the term and most of what comes back is cyber security or AI testing, where a red team attacks a network or a model to find its weaknesses. We apply the same instinct to plans and decisions. Having a red team avoids the risk of a room full of people all thinking along the same lines. In a wargame the red team is usually the side playing the opposition: with somebody actively trying to prevent your plan from succeeding, you can see where an opponent or competitor will seek to attack you.

Idealised thinking

Linear planning, mapping a route from A to B without testing it against an unwilling reality, fosters overconfidence and groupthink. The plan looks coherent because no one has tested it; then contact with reality breaks it and the failure cascades. Idealised thinking is the dominant failure mode in well-resourced organisations with confident leadership, not a marginal one. The planning fallacy is its best-known symptom. The corrective is to put the plan under intelligent stress before the world does.

Teams and learning

Shared mental models

Eduardo Salas and his colleagues defined a shared mental model as the knowledge the members of a team hold in common, which lets each of them explain what is happening and expect what comes next, and so coordinate and adapt without having to talk it through. His later work with Scott Tannenbaum (Teams That Work, 2020) breaks that common knowledge into eight types, from vision, priorities and roles to who holds which expertise and what to do if something specific happens. It is what people mean by being on the same page. Tannenbaum and Salas name scenario-based training and debriefs among the ways to build it, which is why we recommend running the same team through a progression of exercise scenarios.

Closed-loop communication

Closed-loop communication, as Eduardo Salas and colleagues define it (McIntyre and Salas, 1995), has three steps: the sender starts the message; the receiver takes it in, interprets it and acknowledges it; and the sender follows up to make sure what arrived is what was meant. It is the readback on a radio net, or between a pilot and air traffic control. In Salas's "Big Five" model of teamwork (Salas, Sims and Burke, 2005) it is one of three coordinating mechanisms, alongside shared mental models and mutual trust, that hold the rest of teamwork together. Without the third step, the team is acting on the sender's hope that the message landed. Because it can be seen and heard, it is easy to practise and easy to spot in a debrief, and an exercise under time pressure is where its absence shows first.

Experiential learning

Experiential learning is, at its simplest, learning by doing, reinforced by reflection that lets people make sense of what they did. The model most practitioners use is Kolb's experiential learning cycle (David Kolb, 1984), which sets out four stages in a loop: a concrete experience, reflective observation of it, abstract conceptualisation that turns the reflection into a principle, and active experimentation that tests the principle next time round. The complete cycle is what makes it work. An afternoon of go-karting supplies the experience and stops there; a scenario exercise runs the whole loop through the debrief, and goes round again with the next scenario.

The debrief

The debrief is where the learning in an exercise happens. In Kolb's cycle, the exercise supplies the concrete experience and the debrief supplies the reflection and the principle drawn from it. Skip the reflection and you are left with an activity. Scott Tannenbaum and Christopher Cerasoli's meta-analysis of 46 samples (2013) found that debriefs improved performance by about 25% on average. We run two. The Hot Wash is team-led, in the minutes after the exercise ends. The After Action Review follows and turns what was noticed into actions with owners. Tannenbaum and Cerasoli found the effect was stronger when the debrief was facilitated and structured, which is why the After Action Review has a facilitator: someone outside the team asking the questions, so the discussion reaches the decisions people would rather not revisit.

Iterative learning (Duffer's Drift)

Ernest Swinton's The Defence of Duffer's Drift (1904) is the original case for iterative learning: a young officer tries to hold a river crossing, fails, learns and tries again across six dreams until he holds it. Competence is built the way strength is, in sets and reps. A rep is one run at a decision; a set is a block of runs on the same skill, with a debrief after each. Anders Ericsson's research on deliberate practice (1993) agrees: expertise comes from practice designed to work on a specific weakness, with immediate feedback and repetition, not from experience alone. For a team, that means a programme of scenarios, each building on the last.

Emergent leadership

Emergent leadership is leadership from someone with no formal authority: the team member the others come to see as exerting leader-like influence, the definition a recent review of thirty years of research settled on (Hanna, Smith, Kirkman and Griffin, 2021). The British Army has been looking for it since 1942, when the first War Office Selection Board set candidates leaderless group tasks: no leader appointed, a problem to solve, and observers watching how each candidate balanced doing well themselves against supporting the rest of the group. Today's Army Officer Selection Board is its descendant. A scenario exercise creates the same conditions. The participant who steps in when the team is stuck, or asks the question nobody else will, shows up leadership that normal operations rarely reveal. The debrief lets them see it too.

Inclusive facilitation

Inclusive facilitation is how we make sure every voice is heard during an exercise. For the quieter members of a team, the exercise is a safe place to speak up. We put the crucial details of the scenario with the specialists who would hold them in the real world, so leaders have to learn to ask the right questions. Confidence to communicate can be built this way. For the extroverts, the same structure is practice in listening: they have to wait, hear the specialist out and understand the technical detail before framing decisions.

Why fictional scenarios work

A scenario that is fictional, occasionally absurd, or set outside the team's own sector can produce sharper learning than a realistic one, because it strips out the baggage: who said what last quarter, whose budget is at stake, the politics of disagreement. People reason about the problem in front of them, and the learning transfers back. The distinction is between realism and relevance; a scenario needs the second and only sometimes the first.