Strategic velocity

The fastest route to greater velocity and agility is to interrogate the strategy process already running and the strategy it produced. Begin by reverse-engineering the current strategy into explicit choices. What is the guiding theory of winning customers over competition? Where has the company chosen to play, and where has it chosen not to? How does it intend to win in each of those places? What capabilities and management systems did those choices commit it to build? Leadership teams often find this harder than expected, because a significant share of what is labelled as strategy turns out to be a list of aspirations rather than a set of distinct choices about what to do and, more consequentially, what not to do.

The second step is the one almost nobody takes. Once the current strategy is clearly mapped, articulate what would have to hold true for the strategy to work. Start at the overall level – and dig deeper into critical strategic bets. Assumptions about markets, customers, and their evolving needs. About competitors, direct, indirect, and substituting. About collaborators across the value chain. About the political, economic, technological, and regulatory context. And about the company itself, its capabilities and its cost position. Map them and assess these assumptions by their potential impact and uncertainty. The high-impact and high-uncertainty bucket is the real strategic agenda that should keep you awake. The rest should keep you on your toes.

The third step is to test the critical assumptions against data and hard facts, and to do so at arm's length. Start simple with AI – but do not rely on data from the past. You also need new data. Apply rigour in test design to avoid biases. The results change the character of the conversation in a way that is difficult to overstate. When assumptions are made explicit and tested independently, making strategic choices ceases to be a contest of conviction among executives and becomes a matter of assessing evidence. It also minimises latent conflict, because the team agrees on what constitutes a valid test before it sees the result.

A global healthcare company applied this logic to its global sustainability strategy. Rather than starting from a blank page, the leadership team mapped the choices it had already made, surfaced the assumptions those choices rested on, including assumptions about supplier readiness and the pace of the renewable energy transition, and tested them systematically. Several held. The ones that did not pointed directly at the supply chain, where the overwhelming majority of the company's total emissions actually originated. That reframing produced commitments considerably more ambitious than the original strategy, and considerably better grounded in where the impact sat.

Where tests come back red, the response should be a short, focused strategy sprint – often run by a small cross-functional team with a clear mandate: Frame the issue, create genuine options, prototype and test them, decide. Framing the issue is the step most often skipped, and it determines everything downstream. The hard part of strategy is rarely choosing among presented options; it is diagnosing which problem is actually load-bearing. The quality of the choices a company makes will never exceed the quality of the issue it framed.

Stress-testing establishes whether a strategy survives the world that can be observed today. However, it says nothing about the worlds that cannot. The default mode for looking ahead is extrapolation: Take the past, project it forward, present it as a forecast. It is comfortable, it survives board scrutiny, and it fails at precisely the moments that matter most.

The Danish market for electric vehicles offers an instructive illustration. Market forecasts from leading analysts published around 2010 and again in 2012 projected steady growth before rapid uptake in the national fleet. Actual adoption stayed well below those curves for many years and then bent sharply upward. The forecasts were roughly five years off the mark, in both directions at different points. A business case built on the projected line would have been wrong twice: first by over-investing into a market that had not arrived, then by under-investing just as it did.

The remedy is not better forecasting, and it is worth being precise about what scenario work does and does not deliver. The empirical record on scenarios improving predictive accuracy is thin. The record on improving preparedness is considerably stronger. Scenarios do not tell an organisation which future will arrive. They ensure that when one does, the response has already been rehearsed, and the trigger has already been agreed upon.

That structure begins with precision about what is being scanned. Key trends are changes already underway. Weak signals are discontinuities that may grow into observable trends. Wild cards are imaginable outliers with low probability and high impact. Black swans are the genuinely unprecedented. Most organisations monitor only the first category, which explains why they are so reliably surprised by the rest. The futures cone from foresight research makes the same point: The projected future, extrapolated from the past, is the narrowest band inside a much wider space of plausible, possible, and potential futures. Corporate planning tends to mistake the narrowest band for the whole. Or at best, an upside and a downside scenario are calculated for the business case with limited clarity about the assumptions driving the differences.

In practice, the work has three parts. Build a clear 3–5-year base case that captures what the organisation believes to be certain. Be clear about those core beliefs that provide firm ground under your feet. Where uncertainty is high, identify the driving forces that are simultaneously most uncertain and most consequential, and build alternative scenarios around them. Tell that story and bring the scenarios to life.

Then take the step most scenario exercises omit. Work future-back. Place the leadership team inside each scenario and ask what the current strategy looks like from there. Which choices still make sense? Which appear naive? What would the team wish it had started two years earlier? That is where real value is created – far beyond an entertaining workshop exercise easily forgotten.

A Northern European player in electric mobility infrastructure did this well. It built four distinct market scenarios around two forces it could not control – the level of political support for electrification and the speed at which vehicle manufacturers entered at scale – and developed four corresponding sets of strategic choices. The value was not in predicting which scenario would arrive. It was that the management team stopped having circular debates about the future and started having precise conversations about trigger points. What would have to be observed to conclude the market had moved into a different world, and what would be done in the weeks following that observation? The result was the postponement of significant investments that if executed had turned into sunk cost.

Those trigger points are where scenario work joins the assumption register and becomes an active part of your strategy monitoring system.

One finding should unsettle anyone who has commissioned a major analysis. In a study of more than a thousand significant business decisions, research found that the quality of the process used to reach a decision explained roughly 53% of the variation in outcomes, while the quantity and detail of analysis explained approximately 8%.³ While the study relied on retrospective self-assessment, so the causal direction is not established, the ratio is wide enough to capture your full attention.

The study is a classic. But the implication sharpens rather than weakens as analytical capacity becomes cheaper. If AI collapses the marginal cost of producing high-quality secondary analysis, the constraint moves further towards the things that were always scarce – framing the right issue, holding productive disagreement, aligning conflicting opinions, and knowing who decides. Organisations that respond to cheaper analysis by producing more of it will find they have optimised the 8%.

Agility, then, is not primarily a process attribute. It is a distributed human capability residing in how several hundred leaders frame issues, weigh options, and commit resources. If every consequential choice must travel upwards to a small group at the top, the organisation will be slow regardless of how elegant its governance model appears on a nice slide.

Three mechanisms make the capability of making better strategic choices scalable.

1. The first is simple rules. When decisions are distributed, a short set of clear, non-negotiable boundary conditions outperforms an elaborate framework, because rules travel across contexts and frameworks rarely do.
   
   
2. The second is clarity about who decides. "Push authority towards the knowledge" and “Let’s empower the team” remain slogans until specific decisions have clear owners. The discipline of assigning who recommends, who provides input, who agrees, who decides and who executes is consistently underdone. Ambiguity about decision rights is the most common and most fixable source of organisational slowness.
   
   
3. The third is a shared language and capability about strategic thinking and decision-making. Choice-making is a muscle. It requires deliberate training and practice, a common approach and understanding what good looks like. A single toolkit launch achieves exactly nothing.

A global market-leading logistics company treated this as a capability programme rather than an event. Instead of another leadership offsite, it codified how the company solves problems into three durable assets: a small set of core principles, a scalable toolbox, and one agile process for moving from issue to decision to impact. It then trained that capability well beyond the executive team, into the management layers where most consequential decisions were being made. The toolbox mattered less than the outcome. Several thousand people came to share one way of approaching strategic choice-making. And the ways of working were changed at scale. Sometimes hardwired into governance models – and sometimes just embedded as “how we do things around here.”

Scaling great decision-making asks something specific of senior leaders. The role shifts from being the heroic decision-maker to becoming a social architect: The executive who designs how a choice gets made: someone who ensures the right people and the right dissent are present, and who resists supplying the answer before the question has been framed. That shift is harder than it sounds, because many organisations continue to promote and reward the opposite behaviour. But leaders who truly understand the classic study mentioned tend to know better.

Individuals cannot behave with agility inside machinery built for annual predictability. At some point the system itself has to be redesigned.

The starting point is a due diligence of how strategy is actually developed and executed today, not how the official process describes it. Where does the system create value, where does it stall, what is the historical success rate of the initiatives it produces, and how many resources moved between units last year? Fascinatingly, most companies have never examined their strategy process with the rigour they routinely apply to a supply chain.

That last question is the sharpest instrument available. A budget correlation near 92% year over year tells an executive team more about its real strategy than any document it has produced. Reallocation is where agility becomes measurable, and it is where most operating systems fail, because the machinery for adding initiatives is well developed and the machinery for stopping them is not. Strategy is as much about unmaking choices as making them. Exit, divestment, and initiative termination are the least practised competencies in most portfolios, and the ones that free the resources every new choice requires.

The design objective is what John Kotter, emeritus professor at Harvard Business School, called a dual operating system: a hierarchy that delivers reliably alongside a faster network that acts on opportunity. On one side, the periodic ritual becomes lighter, simpler, and considerably better at forcing genuine choices rather than producing glossy consensus documents. On the other, responsibility and authority to act on emerging opportunities are distributed widely.4

Five components should be deliberately designed. The strategy cascade, which should work as a dialogue-based choice-making process in which every level frames and makes real choices within the frame above it, rather than receiving strategy as passive recipients. Communication and sensemaking, which are the means by which a strategy becomes broadly shared, embedded, and personally meaningful. Essentially, each leader and employee should be able to answer why strategy matters and what choices it demands. Strategy governance, which drives alignment and performance management, oversees the assumptions and the initiative portfolio, and enables fast redeployment of capital, talent, and time. The impact initiative engine, where small teams work to a shared delivery method with behavioural as well as business impact targets, because observable behavioural change is the earliest indicator that strategy is actually creating impact. And the leadership principles, which turn the model from governance into practice: unity and accountability in the leadership team and leaders who walk the talk rather than sell the slides at another townhall.

A Nordic energy group has been building exactly this machinery across three commercial business units. The units serve overlapping customers through different products, which gave the group the classic integration problem. Each unit could optimise itself perfectly and the group could still lose, because the customer crossovers, the channel conflicts, and the capital trade-offs all sat between the units rather than inside any of them.

The design has three tiers of strategy forum: group level across the units, one inside each business unit, and one inside each business area. What makes it an operating system rather than three parallel meeting series is that all of them are synchronised into a single recurring Strategy Week, held quarterly, with lighter pulse meetings in between. The week opens at group level looking outward, through a shared external radar, the assumption register, and strategic questions prepared in advance. It cascades down through units and areas. It closes back at group level looking inward, at indicators, cross-unit initiatives, the strategic option board, and rolling financial planning.

What that rhythm creates is not just more meetings on top. Actually, it creates less. It is a standing forum with both the information and the mandate to move resources between units frequently, and a visible record of both assumptions that need monitoring and which options are active, which are paused, and which remain open. Cross-unit forums that lack that mandate produce coordination and more meetings. This one is designed to produce clear-cut choices and agility.

Strategy in the age of AI

Until recently, the scarce resource in strategy was often deep analysis. In the age of AI, that assumption is expiring.

Testing assumptions becomes continuous, making it easier to monitor the assumption register. Agentic systems can monitor signals, investigate what they mean for strategic priorities, and trigger follow-up analysis, even without being asked.

Scenario and futures thinking work becomes much cheaper to produce and therefore easier to overproduce. The discipline shifts from building scenarios to selecting which uncertainties deserve the compute and token spend. Generating is cheap. Evaluating is hard.

Choice-making now genuinely can scale. When frontline teams have analytical capability at their fingertips, decisions can move to where the knowledge sits, provided someone has defined the guardrails and escalation criteria that keep the whole thing coherent.

Moreover, cheaper monitoring removes the excuse for not reallocating resources. When a draft case for entering a market – or leaving it – can be produced in the same afternoon, the organisation's ability to move resources quickly becomes key to success.

While AI is disrupting what might be labelled the easy parts of analysis, it is worth remembering that AI does not produce new knowledge. To truly discover what is next, you need real primary research and novel insights. And to facilitate better strategic decisions that are deeply anchored and well-aligned, you still need capable humans and real conversations in the loop.

Making strategy work in a turbulent world

Strategy is not dead. What is dead, however, is the assumption that a good strategy can be produced once and executed in an unchanged form for five years. The world has moved and is constantly moving. Agility, properly understood, is not the ability to change everything quickly, but knowing which commitments to hold tight through turbulence and which to hold loosely – and having the instruments and operating system to tell the difference in real time.

The four moves run at different clock speeds. A stress test of the current strategy can be completed fast. Painting and understanding pictures of the future take longer. Building choice-making capability and redesigning the operating system are journeys. There is no requirement to sequence them neatly: Start wherever the pressure is highest.

One caution against the obvious temptation: The next level of strategy should not be about adding complexity, building more advanced simulations, or validating every fact through exhaustive analysis. On the contrary, it should simplify the process. Engage the organisation properly. Get to decisive action on clear-cut choices. And keep a live register of what would have to be true for those choices to work, so that you learn the ground is shifting from your upgraded strategy instruments, rather than from results that surprise you because they fell outside your five-year plan.