What kind of strategy framework this is
The conventional strategy framework answers the question: what move should we make?
It picks among options. It optimises for advantage. It assumes the environment holds still long enough for the move to land.
The environments most of my clients operate in do not hold still. A founder finds an ICP that worked last quarter and finds it cold this one. A scaleup is told by its accelerator that it must become a product company by Tuesday. A corporate operator inherits a portfolio company whose binding constraint moved overnight when a foundation model shifted the cost curve underneath it. The conventional question, "what move should we make?", arrives late.
By the time you have answered it, the constraint has rotated.
Critical Path Layers is built for the prior question. Given that the environment is moving, given that you have already chosen what to work on, given that your resources are finite, which layer of work is load-bearing right now? It is a framework for sequencing under speed and uncertainty. It is diagnostic before it is prescriptive. It tells you which thing to solve first because solving anything else is, in expectation, a waste of the runway you have.
That is the kind of framework it is. The rest of this piece places it within the genealogy of strategy thinkers who have grappled with the same problem from different directions, names the specific debts CPL carries to each, and identifies the line CPL walks that none of them quite walks.
The architecture, briefly
CPL operates on five process layers. The layers describe the diagnostic work, not the content of any particular initiative. They sequence as follows:
Situational awareness. What has moved, what is settling, where the dependencies have shifted.
Dependency mapping. What depends on what; the structural read of the system.
Binding constraint. Of all the dependencies, which is load-bearing right now.
Gated sequence. The order of work that resolves the constraint, with explicit gate criteria.
Constraint rotation. Watching for the signal that the situational read has gone stale and a re-entry at Layer 1 is needed.
The references that follow trace what each thinker contributes to the architecture above. Where contributions land at a specific process layer, the layer is named.
Click on the image to access to Interactive Visualisation
The strategy lineage CPL sits inside
Nine thinkers and traditions are load-bearing for what CPL does. Two more sit nearby and are worth naming as boundary cases.
-
Wardley's contribution is the most direct intellectual neighbour. His maps treat strategy as a function of situational awareness rather than choice from a menu. They make the landscape visible: where you are, what is evolving, what is settling into commodity, what is your inertia. The sixteen forms of inertia he names are the most useful taxonomy of organisational resistance written in the last twenty years.
CPL borrows three things from Wardley. The diagnostic-first stance: look at the terrain before you choose a move. The co-evolution test: practices and components shift together, so a strategy that does not account for what is moving around the move is brittle. And the explicit naming of inertia as a system property, not a character flaw of the people resisting.
Where CPL differs: Wardley maps the landscape; CPL sequences the work. A Wardley map shows you what is genesis, custom-built, product, and commodity. A CPL diagnosis shows you which dependency is load-bearing for the next ninety days. Wardley feeds Layer 1 (situational awareness) directly. The mapping discipline is, in many ways, the practitioner-grade instrument Layer 1 currently lacks; CPL's own Layer 1 instrument, in development, borrows structure from Wardley.
-
McGrath's work begins from the premise that durable competitive advantage is no longer the central planning assumption. Advantages are transient. The strategic task is not to defend a position but to design the cadence of exit, reinvention, and re-entry that keeps the portfolio of advantages refreshed. Her discovery-driven planning method makes the assumption set explicit and revisits it at gates, not annually.
CPL shares the assumption that advantage is unstable. It also shares the gating logic: each process layer has criteria that must be credibly met before downstream layers become productive. Where McGrath operates at the portfolio of advantages and the cadence of reinvention, CPL operates one altitude down: at the architecture of work inside a single initiative. The two frameworks compose. McGrath tells you when the advantage you hold is decaying. CPL tells you which layer of work must clear first before the next advantage can stand up. McGrath's gating logic informs Layer 4 (gated sequence) directly.
-
Reeves' central claim is that strategy frameworks must vary by context. Classical strategy works for predictable, malleable environments. Adaptive strategy for unpredictable but malleable ones. Shaping strategy for environments where you can change the rules. Visionary for situations where you must build a new market. Renewal for crisis. The mistake is to apply one school across all contexts.
CPL shares the meta-claim that frameworks must vary by context. It expresses the variance differently: not by selecting which school of strategy to apply, but by identifying which process layer the current diagnostic work demands. The "which school" question and the "which layer of work" question are siblings, both grounded in the refusal to flatten strategy into a single recipe.
Where CPL adds value next to Reeves: the layer model is auditable. You can ask whether a given layer's gate criteria are met. You can ask whether the work being done belongs in this layer. You cannot ask, in the same direct way, whether a strategy is correctly classified as adaptive or shaping. The classification matters but the test is harder.
-
Christensen's two contributions are widely cited and rarely fully absorbed. Jobs to be Done reframes the customer question from "who is the buyer?" to "what job is the buyer hiring this product to do?" The Innovator's Dilemma names the structural reason incumbents fail: they optimise for their current customers and miss the markets where worse-on-current-metrics products become good-enough-on-future-metrics.
CPL borrows JTBD directly. At Layer 3 (binding constraint), JTBD is the structural lens you apply when the presenting symptom looks like a customer problem. The question is whether the buyer's job-to-be-hired is correctly named. In organisations buying internally, the buyer often hires the initiative for a job-as-narrative ("look forward-leaning at next board pack") rather than a job-to-be-done ("move a P&L line by three points"). The narrative job and the operational job have different gates. The initiative that lands the first will not survive a budget review; the initiative that lands the second will. JTBD is the diagnostic that surfaces which job is actually in play.
Christensen named the customer reframe but did not sequence the work that depends on it. CPL puts the JTBD diagnostic at Layer 3 and refuses to let Layer 4 (gated sequence) become productive until the JTBD is resolved.
The disruption logic operates at Layer 1 (situational awareness) and Layer 5 (constraint rotation). A transformation initiative misclassified as optimisation when the actual phenomenon is structural change is the same error Christensen identified at the incumbent level, applied inside a single organisation, surfaced when the situational read goes stale and the constraint rotates beneath the original frame.
-
Martin's "presumption of guilt" piece names a structural fact that most strategy work ignores. The status quo never has to prove itself. It is embedded. Any proposed alternative bears the full evidentiary burden. That asymmetry is not bias; it is governance. It is also the reason most transformation initiatives die quietly: they are evaluated against a standard the incumbent option never had to meet.
CPL inherits this insight in two places. In the organisational immune response (a cross-cutting dynamic that recurs regardless of layer), the four mechanisms (procedural resistance, resource competition, standard dilution, passive waiting) are the operating expressions of Martin's asymmetric burden. At Layer 3 (binding constraint), the same asymmetry shows up as a pilot that succeeds technically but does not clear the buyer's organisational burden of proof, because the burden is heavier than the operator was told.
Martin's strategic choice cascade (what is our winning aspiration, where do we play, how do we win, what capabilities, what management systems) is structurally compatible with CPL but operates at a different altitude. Martin sequences strategic decisions. CPL sequences strategic work. The difference matters: Martin tells you to decide where to play; CPL tells you that until your problem-solution fit is real, your "where to play" decision is a hypothesis with no test built around it
-
The Other Side of Innovation makes a claim that is structurally invisible until it is named: the existing organisation is doing exactly what it should be doing to keep the performance engine running, and that is what kills transformation initiatives. Resistance is not the wrong people in the wrong roles. Resistance is the right people in the right roles operating in the wrong context.
CPL builds on this directly in the immune response dynamic. The four mechanisms are not "what bad organisations do to good initiatives". They are what governance functions do, correctly, when asked to evaluate something they were not designed to evaluate. The reframe matters because it changes the intervention. You do not fight the immune response; you classify the initiative correctly so the right governance is applied. That classification happens at Layer 3 (binding constraint): the diagnostic question is whether the binding constraint is the initiative's design or the receiving organisation's governance fit.
-
Keeley's Ten Types of Innovation is empirical strategy. The argument: single-dimension innovations are the easiest for incumbents to copy or neutralise. Innovations that cluster across multiple dimensions (offering, configuration, experience) are harder to dislodge. The case is built on a large dataset, not a thought experiment.
CPL borrows the clustering insight but uses it differently. In Wardley, Martin, and Govindarajan terms, Keeley's clustering shows why incumbents push initiatives toward product (the most defensible terrain for them) and away from model and process (the terrain where the initiative could actually win). The immune response operates on this: it accepts product-layer initiatives and rejects model-layer ones, not because of malice, but because the organisation's existing capabilities can absorb the former and not the latter. Layer 3 surfaces this as a dependency check: is the binding constraint at the product layer or at the model layer?
-
This is the debt that has been least visible in the published writing. Meda's work, spanning a decade from the Focused Innovation Canvas (2016) through his recent writing on problem dynamics, makes a claim few framework-builders make explicitly: the building blocks of a business model are not boxes to fill but a directed graph to traverse. Problem connects to Added Value connects to Product in a specific sequence. Optimise an element that depends on an unvalidated upstream element and the optimisation is wasted.
CPL applies that same logic at the organisational level. Meda asks whether the business model holds together. CPL asks whether the organisation is working on the right layer of strategic process at the right time. The dependency frame is shared. The unit of analysis is different: Meda's nodes are business model elements; CPL's nodes are diagnostic acts. This is the conceptual debt CPL carries most heavily and the one most worth naming publicly. Meda's directed-graph claim is structurally adjacent to CPL's Layer 2 (dependency mapping).
-
Jackson's Four Levels of Product-Market Fit (Nascent, Developing, Strong, Extreme) does for PMF what CPL does for the broader arc of strategic work: gates between levels, criteria at each gate, no premature passage. CPL's Layer 4 (gated sequence) borrows directly from this. Where Jackson stops at PMF, CPL extends the gating logic to the upstream diagnostic acts (situational awareness, dependency mapping, binding constraint identification) and to downstream movement detection.
-
Snowden's Cynefin framework distinguishes complicated systems (which decompose into independently solvable parts) from complex systems (which do not). The distinction is the conceptual root of why dependency thinking is necessary in the first place. A strategic initiative or a transformation programme is a complex system, which means the components are not separable: the ICP work and the pricing work and the pilot work mutually shape each other and cannot be sequenced as if they were complicated subroutines.
CPL takes Cynefin's classification as given, then asks the next question Snowden does not answer directly: given that this is complex, which dependency is load-bearing right now, and how do I know? Layer 1 (situational awareness) and Layer 3 (binding constraint) are the layers most directly shaped by Snowden's framing. The complicated-versus-complex distinction also explains why domain thinking dominates so much strategy writing. Domain thinking treats strategic work as complicated. Dependency thinking treats it as complex. The two cannot both be right; CPL is built on the second.
What CPL claims newly
Reading this lineage, the natural question is: given that the parts are inherited, what does CPL itself contribute? Three claims.
The first claim is unit of analysis. Strategy frameworks operate on positions, options, advantages, choices. CPL operates on layers of work. A layer is not a position you hold or an option you select. It is a diagnostic act defined by what it depends on and what depends on it. Sequencing by layer is what dependency thinking looks like once it has been moved from the business model surface (Meda) to the organisational diagnostic surface (CPL). The work is the unit.
The second claim is about the density and self-service nature of the instrument layer. The lineage is not tool-free. Steve Blank shipped customer development as a methodology with attached manuals. Rita McGrath shipped discovery-driven planning with reverse financials and assumption checklists. Larry Keeley shipped Ten Types with diagnostic cards. Wardley shipped mapping as a practitioner discipline. The honest difference is narrower than "tools versus no tools": it is the cadence at which the framework expects the practitioner to run an instrument, and the conditions under which they can run it without the framework's author in the room.
CPL is built so that the diagnostic toolkit converges at Layer 3 and the practitioner can run an instrument alone in twenty minutes. The instruments are not ancillary aids that follow the concept; they are the test surface the framework holds itself to. If a pattern of instrument outputs consistently fails to surface the layer at fault, the framework needs revision, not the instrument. That falsifiability commitment, the framework wagering itself on its instruments, is rare in strategy work and is one of the structural commitments CPL is making.
The third claim is that strategy under speed and uncertainty needs a diagnostic prior, not a prescriptive one. The thinkers above broadly converge on this. Wardley's situational awareness, McGrath's discovery-driven planning, Reeves' context-dependent selection, Martin's choice cascade, Govindarajan's performance-engine reframe, all argue for diagnostic discipline before prescriptive action. CPL is prescriptive about sequence (which layer of diagnostic work comes next) and silent on content (what to actually do at each layer). It tells you which dependency is load-bearing right now and which downstream work will not convert until the upstream gate is cleared. It does not tell you what to do at Layer 3; only that you are at Layer 3, and that the constraint named here governs the gates at Layer 4. The prescription on content belongs to the operator, the founder, the team. The framework's job ends once the layer and its gate are named.
Where CPL falls in the genealogy
If the genealogy of strategy under uncertainty is plotted as a tree, four lines feed into CPL.
The first line is the situational-awareness lineage running through Boyd's OODA loop into Wardley's maps. This is the line that argues strategy is observation before action. It feeds CPL's Layer 1 and Layer 5.
The second is the dynamics-of-advantage lineage running from Christensen's disruption through McGrath's transient advantage. This is the line that argues advantage is not a position to defend but a cadence to manage. It informs CPL's Layer 5 most directly.
The third is the dependency-thinking lineage running from systems thinking (Senge, Argyris) through Meda's business-model decomposition. This is the line that argues structure precedes content: until the dependency graph is understood, none of the optimisations stick. It feeds CPL's Layer 2.
The fourth is the sense-making lineage running through Snowden's Cynefin framework. Snowden's distinction between complicated and complex systems is the conceptual root of why dependency thinking is necessary in the first place. It frames the whole stack.
CPL sits at the intersection of the four. It is observational like the first, cadence-aware like the second, dependency-ordered like the third, and complex-systems-honest like the fourth. It is also instrument-backed, which none of the four are by default.
The line worth walking publicly
There is a temptation in this kind of writing to claim more novelty than is honest. CPL is not a new strategy paradigm. It is a re-ordering of existing strategy insights into a sequenced diagnostic process with instruments attached at the layer that does the verdict work. The novelty is in the synthesis, the unit of analysis (diagnostic acts rather than positions), and the instrument-backed delivery. The novelty is not in the components.
The honest claim is narrower and stronger. CPL is the operating system for sequencing strategy work under speed and uncertainty. It inherits from the lineage above. It refuses to flatten strategy into a recipe. It produces instruments the practitioner can run without the framework's author in the room. That last commitment is the one that makes it useful at the cadence the environment now demands.
The work of placing CPL in this lineage is not for credibility.
It is for clarity. Naming the debts is what lets the new contribution stand on its own.
Further reading
Simon Wardley, Wardley Mapping (open-source book, 2015 onwards). The taxonomy of inertia is in the chapter of the same name.
Rita McGrath, The End of Competitive Advantage (Harvard Business Review Press, 2013), and Discovery-Driven Growth (with MacMillan).
Martin Reeves, Knut Haanaes, Janmejaya Sinha, Your Strategy Needs a Strategy (Harvard Business Review Press, 2015).
Clayton Christensen, The Innovator's Dilemma (Harvard Business Review Press, 1997) and Competing Against Luck (HarperBusiness, 2016) on Jobs to be Done.
Roger Martin, Playing to Win (Harvard Business Review Press, 2013) and "The Presumption of Guilt" (Medium, 2022).
Vijay Govindarajan and Chris Trimble, The Other Side of Innovation (Harvard Business Review Press, 2010).
Larry Keeley et al., Ten Types of Innovation (Wiley, 2013).
Philippe Meda, Innovation Copilots writings, 2016 to present. The 2016 piece "It's Time to Let Go of the Canvas" is the foundational text on the dependency claim.
Todd Jackson, "The 4 Levels of Product-Market Fit," First Round Capital (2024).
Steve Blank, The Four Steps to the Epiphany (2005).
April Dunford, Sales Pitch (2023).
Dave Snowden, the Cynefin framework. The original 2003 paper with Cynthia Kurtz, "The new dynamics of strategy: Sense-making in a complex and complicated world," remains the canonical statement.