Playing to Win: Running the Five Choices as One Decision

Five boxes, filled in order

Picture the slide. Five boxes stacked down the page, filled in at an offsite in the order they appear. Winning aspiration: be the leading provider in our markets. Where to play: our current segments, plus two adjacencies. How to win: customer intimacy, innovation and operational excellence. Capabilities: a list with one entry per function. Management systems: a balanced scorecard. Every box is full. Nothing has been chosen.

The framework on that slide is A.G. Lafley and Roger Martin's. Their book, Playing to Win, is built around five integrated strategic choices, set as questions: What is our winning aspiration? Where will we play? How will we win? What capabilities must we have in place to win? What management systems are required to support our choices ¹? For the definition and the structure, see the Lexicon entry on the Playing to Win cascade. This page is about using it.

The argument of this page: Playing to Win fails when it is filled in like a form, top to bottom, one box at a time. Where to play and how to win are one decision with two names. Generate a few candidate pairs, test what would have to be true for each, choose one, and the other three boxes follow from it. Done that way, the cascade forces choices. Done as a form, it hides their absence behind five tidy headings.

Five full boxes can still mean nothing was chosen.

What each choice is for, and its usual fake

The publisher describes how Lafley, as CEO of Procter & Gamble and in close partnership with Martin, doubled P&G's sales, quadrupled its profits and increased its market value by more than $100 billion in ten years, and says the book uses stories of P&G brands such as Olay, Bounty and Gillette to illustrate the choices ¹. Those are the authors' own results, told by the people who made them, so treat them as an illustration of the method rather than proof of it. The method itself is easy to state and easy to fake.

The right-hand column has one thing in common: nothing in it disappoints anyone. That is not a coincidence. A cascade filled in to keep the room comfortable will produce exactly that column. Martin observes that executives find strategy scary, and that choosing one means making decisions that explicitly cut off possibilities and options ⁴.

The pair at the center

The most common mistake is treating where to play and how to win as two decisions made in sequence. Martin wrote about this after the CEO of a large Australian company, an eager user of the cascade, called him with a strategy development problem ³. His conclusion: no how-to-win choice can be made outside the context of a particular where-to-play choice, so the only productive way to generate options is to consider matched sets of choices ³. Whether to aim broadly or narrowly, or to compete on low cost or differentiation, matters less than that the answers form a matched pair ³.

Why does the sequence fail? Choose the field first and the team will pick the field that looks most attractive in a market study, which is usually the field every rival also finds attractive. Then it has to invent a way to win there, and the only ways left are generic. Choose the way to win first and you get the mirror problem: a capability in search of customers. The pair has to be chosen together because the advantage only exists on a particular field.

• Matched: where to play, independent veterinary clinics within overnight courier range; how to win, results before the clinic's morning appointments. The field is what makes the promise affordable.
• Mismatched: where to play, “all clinics, including national hospital chains”; how to win, the same morning-results promise. National chains buy on price through central procurement, so the promise wins nothing they will pay for.
• Generic: where to play, “growing segments”; how to win, “superior service.” Any competitor could write it, so it is not a choice.

Where to play and how to win are one decision with two names.

Generate possibilities, then ask what would have to be true

Lafley and Martin, with Jan Rivkin and Nicolaj Siggelkow, described a process for choosing among them. Conventional planning, they argue, lacks the careful generation and testing of hypotheses that makes a method scientific ². Their alternative, in outline, is below ². The notes inside each step are ours, from practice. Before step one, agree what decision is being made and why now; a vague question produces vague possibilities.

1. Generate possibilitiesWrite three to five candidate where-to-play and how-to-win pairs, including at least one the team finds uncomfortable. Each should be a full story, not a direction.
2. Ask what would have to be trueFor each possibility, list the conditions that must hold for it to succeed: about customers, the industry, competitors and your own capabilities and costs. Do not ask whether they are true yet.
3. Find the barriersAssess which conditions are least likely to hold. Those are the barriers to each choice.
4. Test the barrier conditionsDesign tests for the key barrier conditions and run them: customer trials, cost models, conversations with channel partners.
5. ChooseReview the test results and pick the possibility with the fewest serious barriers. At this point, the authors note, choosing is simple ².

The authors' example is P&G in the late 1990s, when it wanted to become a major global player in skin care. After testing the barrier conditions for several possibilities, it chose a bold option that, they argue, might never have surfaced in a traditional planning process: reinventing Olay as a prestige-like product sold to mass consumers. They report that the new Olay succeeded beyond expectations ².

The deeper value is social, not analytical. Arguing about whether a strategy is right sets executives against each other. Arguing about what would have to be true, and agreeing a test, lets the skeptic own the test of the condition they doubt. The authors put it as a shift from asking “What is the right answer” to asking “What are the right questions?” ².

A worked example: a pet food maker picks its pair

Notice the order. The aspiration was written last, once the team knew what winning would mean. Notice also what was given up: grocery, which is most of today's revenue, gets run for cash. A cascade that does not produce a loss that size has probably not chosen anything yet.

Capabilities and systems: where the cascade usually stops

Most teams that get the pair right still run out of energy at the bottom two boxes. The capabilities box becomes a list of every function, because leaving one out feels like an insult. The management systems box becomes a description of the current scorecard.

Two tests help. For capabilities: would this capability matter as much under the possibilities you rejected? If yes, it is hygiene, not a distinctive capability, and it does not belong in the box. For management systems: would a rival with a different strategy run the same measures and meetings? If yes, the systems are not supporting your choices; they are supporting your industry. The KPI examples page shows how to pick measures that test the specific bet a strategy makes.

The cascade also repeats down the organization. In Martin's account, choices flow from the top of the company to the front line like a river, with each level setting the context for the choices below it ⁵. A business unit's where to play sits inside the corporate one; a function's choices sit inside the unit's. The corporate, business and functional strategy page covers what each level has to choose.

The fair objections

“It is a consumer-goods framework.” The worked stories come from P&G brands ¹, and a sharp critic can say that branded consumer products are unusually suited to clean where-to-play choices. The answer is that the questions do not depend on the industry. Any organization with limited money and attention chooses which customers to serve and why they would choose it, whether it says so or not. The strategy vs. tactics page shows what happens when that choice is left implicit.

“It fixes a strategy that should emerge.” Some of the best strategies are discovered, not designed. The what-would-have-to-be-true method is the framework's answer: the choice is a bet with stated conditions, and when a condition turns out false, the bet gets reopened. That is closer to learning than to a five-year plan. Martin himself argues that strategy is about making bets, and that planning which feels comfortable is usually not strategy ⁴.

“Capabilities first is safer.” Many teams prefer to start from what the company is good at. Martin's warning about self-referential frameworks applies: they lead managers to design a strategy entirely around what the company controls, when revenue depends on customers ⁴. Start from the field and the customer. Capabilities are what the chosen pair requires, not where the search begins.

Monday morning

Take your current strategy and rewrite it as a single where-to-play and how-to-win pair in two sentences. Then write two rival pairs that a thoughtful competitor, or a new CEO, might argue for instead.

For each of the three, list what would have to be true. Circle the condition you are least sure of for your current pair. If nobody has tested it, that is the most valuable piece of strategy work available to you this quarter.

The five boxes are not the strategy. They are a check that a strategy exists. A cascade earns its place when someone can point at the second and third boxes and say what the company gave up to fill them, and what evidence would make it change its mind.

Questions

What are the five choices in Playing to Win?

Lafley and Martin frame them as questions: What is our winning aspiration? Where will we play? How will we win? What capabilities must we have in place to win? What management systems are required to support our choices? The choices are meant to be integrated, so the answers have to fit and reinforce each other rather than being filled in one at a time.

Why should where to play and how to win be chosen together?

Because an advantage only exists on a particular field. Roger Martin argues that no how-to-win choice can be made outside the context of a specific where-to-play choice, so teams should generate and compare matched sets of the two. Choosing the field first tends to land on the field every rival also wants, leaving only generic ways to win.

What does “what would have to be true” mean in strategy?

It is a step in the process Lafley, Martin, Rivkin and Siggelkow describe: for each strategic possibility, list the conditions that must hold for it to succeed, identify the ones least likely to hold, test those, and choose the possibility with the fewest serious barriers. It turns arguments about the answer into agreement on what to test.

Is Playing to Win only for consumer goods companies?

The book's stories come from Procter & Gamble brands, but the questions are not industry-specific. Any organization with limited money and attention chooses which customers to serve and why they would pick it. The framework makes that choice explicit and testable instead of leaving it to the budget.

Where to go next

• Strategy vs. Tactics: The One-Question Test
• Corporate vs. Business vs. Functional Strategy
• Playing-to-Win Cascade
• Positioning
• Competitive Advantage
• What Is Strategy? A Complete Guide