Playing to Win: The Strategy Cascade as a Forcing Function

This is one of RoadmapOne ’s articles on Strategy for Product Leaders .

Playing to Win is a strategic choice cascade developed by A.G. Lafley and Roger Martin. It consists of five interdependent questions: What is our winning aspiration? Where will we play? How will we win? What capabilities must be in place? What management systems are required? Each answer must cohere with the others, which forces organisations to make strategic commitments rather than produce aspirational statements.

Most strategy documents are shelfware. Written once, presented once, nodded through, filed in a Confluence space nobody has opened since, and then comprehensively ignored as the actual decisions get made by whoever shouts loudest in next week’s ExCo meeting. I’ve seen this in Series B startups with three people in the strategy review; I’ve seen it in FTSE companies with sixty. The pattern is the same: strategy is theatre, and the theatre is the point.

Playing to Win—the five-question cascade developed by former Procter & Gamble CEO A.G. Lafley and strategy professor Roger Martin—is the only widely-used strategy framework I’ve seen that routinely escapes this trap. Not because it produces a better document. It doesn’t. The output of a Playing to Win exercise looks embarrassingly mundane: five short paragraphs, one per question. You’d be forgiven for thinking it was a glorified to-do list.

What Playing to Win does differently is that it forces the conversation. The five questions are arranged in a cascade—each one depends on answering the previous one honestly—so that evasion becomes visible. You cannot tick the boxes without making choices, and the choices have to be coherent, or everyone around the table sees the incoherence. That forcing function is the whole value. The framework itself is incidental.

Sign up (free) to turn strategy choices into squad allocation

What Is Playing to Win?

Playing to Win is a strategic choice cascade with five sequential questions, each of which must be answered before the next one can be answered coherently:

1. What is our winning aspiration? The purpose. What would it mean for this business—or this product—to “win” in its chosen domain?
2. Where will we play? The boundaries. Which markets, segments, geographies, channels, product categories, and stages of the value chain are we choosing to operate in—and which are we explicitly choosing not to operate in?
3. How will we win? The advantage. What gives us a durable right to win in the places we’ve chosen to play? Channel, brand, network effects, switching costs, proprietary data—pick the ones that actually apply.
4. What capabilities must be in place? The competencies. Which internal capabilities are essential to delivering the How-to-Win? If a capability isn’t on this list, it’s a nice-to-have, not a strategic priority.
5. What management systems are required? The governance. How will you measure progress, allocate capital, and hold people accountable against the choices you’ve just made?

That’s it. Five questions. Forty-five minutes to explain. A lifetime to answer properly.

Lafley and Martin codified the cascade based on Lafley’s P&G tenure, during which the framework was used to turn around Olay from a declining mass-market skin-cream brand into a “masstige” category leader. The book, Playing to Win: How Strategy Really Works, has since become the standard strategy text for a generation of operators, investors, and boards. It is cited routinely in PE investment committees, in CEO onboarding, and—increasingly—in product leadership circles who recognise that “feature roadmap” is not a synonym for “strategy”.

The Cascade Is a Forcing Function, Not a Template

Most articles about Playing to Win treat it as a form to fill in. Open a canvas, populate five boxes, celebrate. This is exactly the failure mode the framework was designed to prevent.

The cascade is a thinking scaffold. Its job is to force answers to be coherent with each other: your Where-to-Play has to be consistent with your Winning Aspiration, your How-to-Win has to justify the Where-to-Play, your Capabilities have to enable the How-to-Win, and your Management Systems have to measure all of it. Any incoherence shows up immediately. If your Winning Aspiration says “category leader in UK legaltech” but your Where-to-Play lists seven unrelated geographies, you haven’t done strategy; you’ve written a wishlist. If your How-to-Win is “great customer service” but your Capabilities list doesn’t include building a support organisation capable of delivering it, the cascade is lying to you and you haven’t noticed.

This is why the discussion matters more than the document. A Playing to Win exercise done honestly surfaces strategic debts the organisation has been carrying without admitting to—assumptions that were never tested, customer segments that were never chosen, adjacencies that were never ruled out. The document at the end is a by-product. The value is in the ExCo member who, halfway through the Where-to-Play conversation, says the quiet thing out loud: “We’ve been pretending to serve enterprise for three years and it isn’t working.” The cascade didn’t say that. The cascade made it unsafe not to say it.

This is also why most organisations can’t do Playing to Win in a single away-day. The useful version is iterative, with each round exposing the incoherences of the previous round. Expect two or three passes before the answers stabilise. If you stabilise in one pass, you haven’t done the exercise—you’ve performed it.

Working Through the Five Questions

1. Winning Aspiration

The Winning Aspiration is not a mission statement. Mission statements are aspirational noise. A Winning Aspiration is a specific, checkable statement of what it would mean to win in your chosen domain.

“Make the world better through technology” is not a Winning Aspiration. “Be the default commercial real-estate analytics platform for UK mid-market brokers” is. The test: could you, three years from now, walk into a board meeting and say unambiguously whether you had achieved it or not?

This question sounds easy and is not. The most common failure mode is confusing “winning aspiration” with “financial target”. £10m ARR is not a winning aspiration—it’s a number you hope to reach because of winning. The aspiration is the competitive position that produces the number.

2. Where to Play

Where-to-Play is the choice question. It defines the boundary of the game you are choosing to play, which necessarily defines the games you are choosing not to play.

This is where the famous Ansoff Matrix lives—it is a tool for answering the Where-to-Play question by categorising product/market choices by risk. It is also where the BCG Growth-Share Matrix , the Innovation Ambition Matrix , and Three Horizons of Growth live—all of them are portfolio-level views of the Where-to-Play choices you have already made. None of them is a substitute for the question itself.

The Where-to-Play choice has at least five dimensions: customer segment, geography, product category, channel, and vertical integration stage. A good answer makes explicit choices on all five. A weak answer makes explicit choices on one or two and leaves the rest as “we’ll figure it out”—which is code for “we’ll default to whatever inbound demand shows up”.

3. How to Win

How-to-Win is the advantage question. Given the game you’ve chosen to play, what gives you a durable right to win? Not to compete—to win.

This is the question where most strategy documents go quiet. “We will deliver superior customer experience” is not a How-to-Win answer; it is a wish. A real answer names the specific mechanism: “We will own the integration into the largest incumbent CRM in this vertical, and that integration will be the default switching-cost barrier for competitors.” “We will build the only compliant data-residency offering for regulated buyers in this geography.” “We will acquire distribution at a cost our competitors cannot match because our parent company owns the channel.”

I will return to this question below, because in 2026 it is the question that has changed most dramatically. AI has collapsed the cost of building a solution to near zero. The cost of convincing a customer to buy it is unchanged. This asymmetry has broken several of the classical How-to-Win answers—cost leadership in particular is no longer a moat—and elevated others (channel, brand, proprietary data, switching costs) to primary importance.

4. Capabilities

Capabilities are the specific internal competencies that make your How-to-Win executable. This is the question that connects strategy to the actual organisation.

The critical test: every capability on the list must be directly traceable to an element of the How-to-Win answer. If a capability is on the list but you cannot explain which How-to-Win mechanism it enables, it is decoration. If a How-to-Win mechanism exists but no corresponding capability is listed, you have a strategic debt—you have committed to winning through a mechanism you do not have the capability to execute.

Capabilities are also where product organisations most often deceive themselves. “Product discovery capability” is genuinely load-bearing in a How-to-Win that depends on continuous customer learning. “Engineering excellence” is not a capability; it is an unfalsifiable virtue. Force the specificity.

5. Management Systems

Management Systems are the governance apparatus that keeps the cascade honest over time. This includes the OKR framework you will use to measure progress against the Winning Aspiration, the capital-allocation reviews that keep capacity aligned with the Where-to-Play, the prioritisation framework—typically BRICE or an equivalent—that weights initiatives by their contribution to the How-to-Win, and the cadence of strategic review that forces the cascade to be revisited.

A strategy without management systems is a one-time event. A strategy with management systems is a repeating conversation that the organisation has quarterly, keeping the cascade’s answers alive as the market moves. This is the question the academic articles skip because it is unglamorous; in practice it is the question that determines whether Playing to Win changes behaviour or does not.

Sign up (free) to connect strategic choices to squad capacity

Playing to Win and the Business Case

Here is the point most articles miss: Playing to Win is the scaffold, not the artefact. The artefact produced by strategy is a business case—a Project Initiation Document (PID) —that every ExCo member signs and commits their functional contribution to. The cascade is what you use to structure the business case so the ExCo can tell whether the commitment is coherent.

Concretely, the five questions map to the sections of a well-written PID:

• Winning Aspiration becomes the PID’s problem statement and segment-level benefit articulation: what would it mean for this product to succeed for this specific customer type?
• Where to Play becomes the PID’s scope section: which segments, geographies, and channels are in-scope—and, just as importantly, which are explicitly out-of-scope?
• How to Win becomes the PID’s differentiation argument: why will this specific mechanism produce the forecast revenue and hold up against competitive response?
• Capabilities becomes the PID’s missing-skills section: which capabilities does the organisation not currently have that must be acquired or built for the product to succeed?
• Management Systems becomes the PID’s accountability and measurement section: who owns which outcome, what triggers a review, and how does the commitment carry into the following year’s budget?

This is why I treat Playing to Win as strictly upstream of the roadmap and the OKR cycle . You answer the cascade, you use the answers to structure the PID, the PID gets signed, and only then does anything get allocated onto the roadmap. Organisations that allocate capacity first and rationalise strategy afterwards are doing strategy theatre. The PID is the artefact that makes passive disagreement impossible; the cascade is the thinking scaffold that makes sure the PID’s answers cohere.

The AI-Era Reframe: Why “How to Win” Has Changed

The five questions have held up well for thirteen years. One of them is breaking.

How will we win? was a question with eight or nine viable answers when Lafley and Martin wrote the book. Cost leadership was a live option. Product differentiation—being first to offer a feature—could hold for years. Being marginally better at execution could be durable. All of this rested on an assumption that has quietly stopped being true: that building a competing solution was expensive, slow, and required scarce engineering talent.

In 2026, that assumption is dead. AI has collapsed the cost of producing a credible software product to near zero. Two engineers with current AI tooling can replicate a feature that used to take a team of twelve a quarter. Your competitor can clone your differentiation in weeks, not years. And because the build cost has collapsed, cost leadership has collapsed with it—nobody has a durable cost advantage when everybody’s marginal cost is approaching the same floor.

What has not changed is the cost of convincing a customer to buy. Customer acquisition, trust, credibility, switching costs, channel relationships, and brand equity all still cost what they cost. A startup with a brilliant feature and no distribution loses to an incumbent with a mediocre feature and a signed framework agreement with the customer’s procurement team. Every single time.

This rewrites the How-to-Win answer. The durable answers in 2026 are the ones AI cannot erode:

• Channel. Existing distribution, incumbent integration, embedded positions inside customer workflows. You cannot clone a framework agreement.
• Brand. Trust built over years. Authority in a vertical. The fact that the buyer can justify the purchase to their board because nobody gets fired for buying X.
• Switching costs. Data lock-in, workflow lock-in, integration lock-in. Structural frictions that make the customer’s choice to change more expensive than the marginal improvement the competitor offers.
• Proprietary data. Datasets that the competitor cannot obtain even if they can clone the product. Closed-loop telemetry from a large installed base.
• Network effects. Value that grows with user count. Uncloneable by feature parity.

What is no longer a durable How-to-Win answer: cost leadership, first-mover feature advantage, “better UX”, or “smarter AI”. These are table stakes; they do not win.

For a product leader working through Playing to Win in 2026, this has a concrete implication: if your How-to-Win answer is a feature list, your strategy is broken. Force the question back up the cascade. The durable winning mechanisms are in channel, brand, switching costs, proprietary data, and network effects—not in what you are building. The build is the easy part now.

Portfolio-Level Playing to Win: One Cascade per Product

A mature company does not have one Playing to Win cascade. It has several—one per meaningful product or business unit—and the cascades may point in different directions.

This is the mistake I see most often in PE due-diligence conversations. The management team presents a single corporate Playing to Win, which the board accepts uncritically, but the portfolio of actual products is spread across three or four different strategic positions: a mature cash-generator in a commoditising category, a growth product in land-grab mode, an acquired product still finding fit, and an internal R&D bet not yet validated. These four products have four different Winning Aspirations, four different Where-to-Play answers, four different How-to-Win mechanisms. Forcing them into a single cascade produces strategic nonsense.

The honest version does one cascade per product—or per business unit when products cluster. Each cascade is answered on its own terms. The corporate strategy becomes the portfolio allocation across those cascades: how much capacity to the cash-generator, how much to the growth product, how much to the acquired-but-unproven, how much to the R&D bet. This is where the tagging frameworks earn their keep—Run-Grow-Transform , Three Horizons , the Innovation Ambition Matrix —they are lenses on the portfolio allocation of capacity across the multiple cascades, not substitutes for answering the cascades themselves.

From Cascade to Capacity

The failure mode I keep returning to is strategy-without-operationalisation. You have beautiful Playing to Win answers. You have a signed PID. You have OKRs. And then in the following quarter, the squad grid shows 80% of capacity being spent on maintenance work for last year’s product, 15% on an ad-hoc request from the Chief Revenue Officer’s favourite customer, and 5% on the supposedly strategic initiative the cascade was pointing at.

The cascade does not execute itself. Strategic choices translate to roadmap only when every squad-sprint allocation—the capacity model inside RoadmapOne —is tagged to a cascade answer. What percentage of capacity is serving the Where-to-Play segments we said we cared about? What percentage is building the Capabilities we said we needed? When leadership looks at the answer and flinches, that is the cascade doing its job—surfacing the gap between the strategy and the allocation in terms that cannot be argued with.

The tagging frameworks in the objective tagging directory are the operational bridge between the cascade’s answers and the allocation. Once you have answered Playing to Win, you know which lens matters most: a business in commoditising mode tags by RGT to make sure maintenance is not crowding out growth; a business in land-grab mode tags by customer journey stage or AARRR to make sure funnel mechanics are getting the capacity they need; a business pursuing a specific Where-to-Play segment tags by segment. The cascade tells you which tagging lens matters. The tagging lens tells you whether the allocation is honest.

When Playing to Win Fails

I have watched Playing to Win fail in three distinguishable ways, and they are worth naming because each of them looks like success in the moment.

Framework worship. The team produces a beautiful cascade document and treats the document as the outcome. The conversation the cascade was supposed to force never actually happens; the cascade becomes shelfware with extra steps. You can detect this by asking an ExCo member, six weeks after the exercise, to recite their own team’s How-to-Win answer. If they cannot—and most cannot—the framework was performed, not used.

Cascade without commitment. The team produces a cascade, signs off on it, and then continues to run the organisation the way they were running it before. Nobody is held accountable against the cascade’s answers. Budgets get set the way they always got set. Roadmap allocation continues to reflect last year’s politics. This is the failure mode the business case / PID is specifically designed to prevent: the signed PID makes passive disagreement impossible, where a Playing to Win document alone does not.

Cascade without AI-era update. The team answers the cascade as if they were still in 2015. Their How-to-Win answer is a feature list. Their Capabilities section is about engineering throughput. Their Management Systems are optimised for internal efficiency rather than channel leverage. The cascade looks competent but is attacking the wrong problem—because the question How will we win? has structurally changed, and the framework does not warn you about that on its own.

Each of these failures is survivable individually. Two or three together will produce a strategy exercise that burns a lot of management time and changes nothing.

FAQ

What is Playing to Win about?

Playing to Win: How Strategy Really Works is a 2013 book by A.G. Lafley (former P&G CEO) and Roger Martin (strategy professor) that lays out a practical framework for strategy: the five-question Strategic Choice Cascade. The book argues that strategy is fundamentally about making integrated choices — Winning Aspiration, Where to Play, How to Win, Capabilities, and Management Systems — and that most organisations confuse goal-setting and planning with strategy. The canonical case study is the P&G Olay turnaround during Lafley’s tenure.

What is the cascade model of strategy?

The Strategic Choice Cascade is the five-question structure at the heart of Playing to Win. It is called a cascade because each question constrains the next: Where-to-Play is constrained by Winning Aspiration, How-to-Win by Where-to-Play, Capabilities by How-to-Win, and Management Systems by all four. The cascade forces coherence — you cannot answer later questions without answering earlier ones, and incoherence between answers is immediately visible.

Is Playing to Win a good book?

Yes — it remains the most widely-recommended strategy book for operators, board members, and product leaders, and with good reason. It’s short, practical, and grounded in Lafley’s actual experience running P&G rather than academic theory. The five-question cascade is the most versatile strategy framework in general use, and the book pairs especially well with Rumelt’s Good Strategy Bad Strategy for diagnosis and Helmer’s 7 Powers for moat testing.

What are the 5 questions of Playing to Win?

The five questions are: (1) What is our winning aspiration? (2) Where will we play? (3) How will we win? (4) What capabilities must be in place? (5) What management systems are required? The questions are cascading—each must be answered coherently with the previous one—and together they form Roger Martin and A.G. Lafley’s Strategic Choice Cascade.

What is the Strategic Choice Cascade?

The Strategic Choice Cascade is the sequence of five interdependent questions at the heart of Playing to Win. It is called a cascade because each question’s answer constrains the next: your Where-to-Play must be consistent with your Winning Aspiration; your How-to-Win must be executable in your chosen Where-to-Play; your Capabilities must enable your How-to-Win; and your Management Systems must measure all of it. The cascade is what turns strategy from a list of intentions into a set of coherent, mutually-consistent choices.

Is Playing to Win still relevant in the AI era?

Yes, but one of the five answers has structurally changed. The framework’s structure is as useful as ever, but the How will we win? question has changed in substance: AI has collapsed the cost of building a competing product to near zero, which means cost leadership and first-mover feature advantage are no longer durable How-to-Win answers. The surviving How-to-Win answers in 2026 are channel, brand, switching costs, proprietary data, and network effects. If a Playing to Win exercise arrives at a How-to-Win answer that is a feature list, the answer is wrong.

How is Playing to Win different from a SWOT analysis?

SWOT (Strengths, Weaknesses, Opportunities, Threats) is a situation-assessment tool: it describes where you are. Playing to Win is a choice-making tool: it forces you to decide where you will play and how you will win. SWOT is an input to strategy; Playing to Win is strategy. The frequent failure mode is teams producing a SWOT and mistaking it for a strategy; a SWOT with no corresponding choice cascade is a description, not a direction.

How do I apply Playing to Win to product management?

Answer the cascade separately for each meaningful product or business unit—one cascade per product, not one cascade per company. Use the five answers to structure the business case / PID that your ExCo signs before any roadmap capacity is allocated. Tag each squad-sprint allocation in RoadmapOne against a cascade answer, so that a capacity-allocation view can reveal whether the actual work matches the strategic choices. Review quarterly; expect the cascade to evolve as markets move.

What is a good example of Playing to Win?

The canonical case study is P&G’s Olay turnaround, in which the brand was repositioned from mass-market skin cream to a “masstige” category leader through explicit choices about Where to Play (the masstige segment sitting between mass-market and premium) and How to Win (product performance combined with channel strategy). More useful for product leaders is to run the cascade on their own product and examine whether their current roadmap allocation is consistent with the answers—most discover it is not.

Conclusion

Playing to Win endures because it is the only widely-taught strategy framework built around a forcing function rather than an output. The five questions are a scaffold for a conversation the organisation would otherwise avoid—about what game it is choosing to play, how it plans to win, and which capabilities it will have to build or acquire to make the chosen winning mechanism executable. The document produced is almost beside the point. The value is the conversation, the choices, and the signed commitment those choices force the ExCo to make.

In 2026, the framework remains essential, but one of the five questions has changed substance: How will we win? is no longer answerable with a feature list or a cost-leadership claim. The durable answers now live in channel, brand, switching costs, proprietary data, and network effects—the dimensions AI cannot erode. Product leaders who use Playing to Win as a thinking scaffold for their business case, update the How-to-Win answer for the AI era, run one cascade per product rather than one per company, and tie every cascade answer to the squad-sprint allocation in their capacity model will find that the framework turns strategy from theatre into a management discipline. Product leaders who treat it as a form to fill in will produce exactly what they were producing before: shelfware.

Pick which side of that you want to be on.

Sign up (free) to allocate capacity against your Playing to Win choices