A few years ago, I wrote that Continuous Strategy is the new Strategy Deployment. I had been looking for a better name for a long time. People mishear deployment as pushing something down onto others. Instead, what it means is moving strategy to where people can use it. I’m now finding that I’m using the term Continuous Strategy more and more.
What that post didn’t do was say much about what the practice actually involves. So this is the longer version. The short version is that a strategy needs to stay in contact with the work, and vice versa. Organisations typically design strategic planning cycles to prevent exactly that.
Why Continuous Strategy?
Almost every organisation I work with has a strategy. There is a document somewhere with a bold title and an ambitious vision. The writers wrote it at an offsite, and the priorities made complete sense that day. However, the writers were the people who then knew the least about the year ahead.
Then the year happens.
The strategy and the world it aimed to fit drift apart slowly, and no one notices the change. We label this outcome strategic drift. People commonly attribute it to a failure of perception. Somebody should have been watching more carefully. That explanation is valid, but in a fast-moving organisation I don’t think it is the main part. Instead, drift is what a planning system produces by design. It is a fixed answer, released into a moving question.
Continuous Strategy keeps the strategy in continuous, two-way contact with the work under it. It doesn’t add a layer to the organisation. Flight Levels already named the strategic level and connected it to the flows beneath. Continuous Strategy describes the interactions between them. It provides steering that lets direction change as conditions change instead of waiting for the next planning round.
We got very good at delivery
Over the past decade, organisations have invested heavily in becoming adaptive at the delivery level, to good effect. They measure flow, shape teams deliberately, and pay attention to cognitive load. However, something was quietly lost in the process: the strategy that once gave those teams their direction. The delivery focus wasn’t wrong. Most organisations just haven’t noticed what went with it.
We left the traditional annual cycle remaining, and it consistently fails for structural, not incidental, reasons. Stephen Bungay’s Directed Opportunism names three gaps between intention and outcome. The knowledge gap lies between what people need to know to plan and what people can actually know. The alignment gap lies between what people asked for and what people did. And the effects gap lies between what people did and what resulted. The instinctive fix for each one is that we should be better at whatever produced it. Plan in more detail. Communicate more forcefully. Measure more tightly. Each response widens the gap we intended to close, and that is what makes the cycle self-reinforcing rather than merely disappointing.
The obvious objection from a board is that the organisation has survived all this before, and that objection is right more often than we might want to admit. However, a long survival record demonstrates robustness, the ability not to fail. That is a different property from resilience, which is the ability to fail safely and recover. Robustness holds until it doesn’t. It also suppresses the small failures that are the only affordable way to learn. The answerable question isn’t how long the robustness window stays open before failure. It’s whether the organisation would notice it closing, and Continuous Strategy is how it would notice.
Sensing without responding
Two models get close to the solution, and each stops short in its own way.
Team Topologies treats team shape as something you design on purpose, and takes the connections between teams as seriously as the teams themselves. Its authors describe it as a mechanism that senses when a strategy must change. A well-shaped topology is an organisation wired to notice: the stream team that catches a feature going unused, the platform group feeling a new kind of demand arrive. Several of its structural decisions are strategic choices in disguise, starting with fast flow itself, which is a bet that in your market the ability to change quickly is worth more than the ability to be right the first time. But a sensing mechanism raises a question without answering it. Sensing runs continuously. Responding strategically, in most organisations, waits for the annual cycle.
Flight Levels does what Team Topologies didn’t. It names the strategic level, connects it to the flow, gives it an interface in Klaus Leopold’s SOFI, and runs continuously. Its limit is precise and easy to miss. There is a cadence for reviewing the work on the strategic board, and no mechanism for reviewing the strategy itself. Adding another meeting wouldn’t fix that, since regular reviews at every level are already part of the model. The issue is that a rhythm built to move work along won’t, on its own, ask whether the direction behind it still holds. Naming a level is not animating it.
What the conversation leaves out
A strategy built by a dozen people looking from the same height will be coherent, and it will not be surprising. Diversity of read is the raw material, and the people closest to customers and operations hold weak signals that no view from the top can reconstruct. David Marquet’s distinction between red work and blue work, in Leadership is Language, explains why they are so rarely asked. The default organisation reserves the deciding for a few people at the top, and gives everyone else the doing. Widening who takes part isn’t consultation after the fact, and it isn’t inclusion for its own sake. It’s forming the direction with the people who can see what the centre can’t.
Something similar happens to the work itself. Every organisation runs several kinds at once, and Explore, Expand and Extract are Kent Beck’s plain names for them. Each needs a different yardstick, and most reviews apply one, in good faith, to all three. Behind the scoring sits a decision nobody makes on purpose: how finite capacity gets allocated across the three. Extract has gravity and takes what it needs. The result is an organisation that keeps getting better at what it already does, while defunding, on schedule, the only work that was going to change that. It’s the same trap I described as strategic insouciance, arriving by a different route. There the mechanism is attention, and focus narrows what an organisation can notice. Here it is capacity, and the funding goes where the returns are already proven. Either way, nothing new gets the room to emerge.
Continuous Strategy as a capability
If the annual document isn’t the answer, a better document won’t be either. What helps is treating strategy as something the organisation practices and improves, in the way it already holds a delivery capability.
That means it isn’t a decision either, and unlike a document it can’t be specified up front. More rigour wouldn’t have saved the plan, because complicated problems have right answers that expertise can find and strategic situations don’t. Anyone who has spent time with Cynefin will recognise the distinction, and will also recognise how often it gets ignored under pressure.
What separates steering from drifting is empiricism, and in Deming’s PDSA cycle the step organisations fail is the adjusting. Plan and do are easy. Checking quietly gets substituted for studying, which is what the shift from PDCA to PDSA was meant to prevent, and it leaves the acting (or adjusting) with nothing much to work from. Even where there is something to work from, adjusting usually means changing the strategy rather than the work, and almost nobody has permission to do that. Continuous Strategy, seen this way, is nested experimentation, where every Do has its own cycle inside it.
There are ten qualities that describe Continuous Strategy when it’s healthy, arranged in five pairs: intent and context, focus and coherence, quantification and decisions, learning and emergence, participation and durability. They are a diagnostic lens rather than a maturity model. There is no score to reach.
The rhythms that make it Continuous Strategy
Signals arriving into an organisation with no rhythm to receive them are read, appreciated, and lost. The Continuous Strategy operating model is the set of rhythms, groups and artefacts that keep intent and reality in contact, and I describe it as four loops separated by scope, each turning at the rate its own feedback arrives.
Planning is annual, and sets the diagnosis and direction. The annual loop itself isn’t the trap, but relying on it alone is. Steering is quarterly and load-bearing, because it is the only loop that runs on the strategy rather than on the work beneath it. Reviewing is monthly, and belongs to the people running the tactics. Refreshing is close to continuous and often isn’t a meeting at all.
Feeding those loops is the harder problem of reading the situation. Two people can look at the same signal and reach opposite conclusions, both credibly, with nothing in the room that either can point at. Cynefin answers what kind of situation you are standing in, domain by domain, with the constraint in each and how to act in it. Wardley Mapping answers where you stand: an anchor, a chain of dependency, an axis of evolution, and the movement most likely to catch an organisation out, which will be the one happening underneath it. Estuarine Mapping offers a third read again on affordances. These are outside-in and inside-out views that most organisations already have somewhere, and never hold against each other.
Making the choices visible
A typical strategy document contains ambitions, activities and KPIs, with nothing in between saying how one produces the other, and it reads perfectly well. Two things are missing.
The first is the guiding policies, written as choices like even-over statements so that each one gives something up. The second is the relationships, which no list can hold. The X-Matrix puts True North, Aspirations, Strategies, Tactics, and Evidence on one page and asks, cell by cell, what correlates with what. The reading is the point. A strategy with no correlations to any tactics might indicate an intention nobody is funding. A tactic with no correlations to any strategy might indicate either a policy nobody wrote down, or work that belongs to a different conversation entirely.
Passing it around
A page built by six people and sent to four hundred becomes an instruction through the ordinary work of tidying it up. Catchball is what prevents that: a consent process for translating strategy into daily work, played by passing artefacts between the people it affects. Consent rather than consensus, because chasing agreement from everybody dilutes a proposal until it has assent and no commitment, while consent obliges a dissenter to improve it. Two A3s are the ball. The Backbriefing A3 is where a team says back what it heard and what it proposes to do about it, which catches leadership’s curse of knowledge as often as it catches the team’s misreading. The Experiment A3 is where a tactic states what it is testing, and what success and failure would look like.
None of this happens at a single level, or in a single meeting. A strategy becomes continuous in the traffic between events. Three groups collaborate in pairs, and the three things each pair works out are Bungay’s gaps again. What couples two cadences is people rather than documents, because a report holds conclusions, while somebody who was in the room carries context, such as why a decision was made and who was unconvinced. What is important is that feedback from the work reaches the strategy level directly, without consolidation at each level rounding off the interesting part.
No finished state
A Continuous Strategy practice decays as soon as it stops being attended to. That is an unsatisfying thing to end on, and also the honest one. Organisations never finish implementing Continuous Strategy.
However, what it buys is a shorter interval between something important being found out and something being decided about it. It also buys a guiding policy that can move without anybody having to call it a failure, and the ability to answer a challenge that the last diagnosis never addressed.
So the ask isn’t to design an improved future operating model. It’s to do the smallest thing you can today. Whatever practice, quality, or cadence you would eventually like to have, the only route to it runs through the adjacent possible, which is whatever is reachable from where you already stand. That is usually smaller and less impressive than anybody wants it to be. Make the move that is available now, and the next one becomes available too, and so on.