Strategy Is a Series of Bets. Most Decks Hide That.

The strategy deck arrives on a Thursday. It is sixty-two slides. It has been worked on for the better part of a quarter by three people who are, individually, very good. The cover slide is titled "FY26 Strategic Plan." The first substantive page is headed, in confident sans-serif, Where We Will Play. The next is How We Will Win. By slide thirty, a five-year revenue waterfall climbs from $480M to $1.2B with the smooth, faintly menacing geometry of something that has been line-fitted to a target.

Nowhere in the document does the word probability appear. Nowhere is there a number with a range around it. Nowhere is there a sentence of the form we are betting that X, at roughly Y odds, and if Z happens we will know within six months that we were wrong. The deck reads as a forecast. It should read as a portfolio of bets.

This gap — between strategy as it is written and strategy as it actually is — is the single most consequential failure of language in modern corporate practice. We have watched it cost real companies real billions. We have also watched the fix, which is unglamorous and almost embarrassingly simple, transform the conversation a board has with its CEO.

The language problem

Strategy decks are written in the declarative mood. We will enter the European market in H2. We will reposition the mid-tier product. We will acquire two regional distributors and integrate them into a single platform by end of year. The grammar is the grammar of certainty. Each clause is a claim about the future, indistinguishable in its tone from a claim about the past.

The reality is different. The European entry has, perhaps, a sixty percent chance of clearing its first-year revenue threshold, conditional on hiring a country manager who has not yet been identified. The repositioning is a bet that a brand association the team has spent a decade building can be re-pointed in eighteen months — a feat with roughly two known analogs in the industry, one of which worked. The acquisition integration is the kind of thing that, industry-wide, hits its synergy targets about thirty percent of the time and overshoots its timeline by a factor of one and a half on average.

The deck does not say any of this. The deck cannot say it without violating the rhetorical contract executives have learned to honor: strategy is the act of choosing, and choices are spoken in the declarative. To say "we will enter Europe at sixty percent confidence" is to sound, in the room, like someone who has not yet made up his mind. To say "we will enter Europe" is to sound like someone who has.

Roger Martin's framing in Playing to Win helps here: strategy is a set of integrated choices about where to play and how to win. It does not help enough. Choices, in the Martin sense, are still spoken as declarations. The probability that any given choice was the right one is left implicit — which means it is left ungoverned.

Why this matters more than it appears to

The cost of declarative strategy is not that the strategy is wrong. Sometimes it is wrong; sometimes it is right. The cost is that the company cannot tell the difference, in advance, between two strategies that the board should be evaluating very differently.

Consider two plans, both presented to the same board in the same quarter, both producing an identical five-year expected revenue of roughly $900M.

Plan A is a single, concentrated bet on a new product line for an existing customer base. The team has built three before and shipped two. The customer demand signal is strong. The execution risk is concentrated in one cross-functional integration that, if it works, produces $900M plus or minus $80M. If it fails, it produces about $300M and a one-time write-down.

Plan B is a four-part portfolio: a modest expansion of the core, a small acquisition, a geographic move, and a price increase. Each component is independently weaker than Plan A's central bet, but the components are uncorrelated. The expected value is the same $900M. The range, because of the diversification, is narrower — roughly $900M plus or minus $40M, with the downside floor closer to $700M.

A board reading the two decks side by side, with neither deck mentioning variance, sees two plans that produce $900M. It will likely pick Plan A, because Plan A has the cleaner story, the bigger headline number, the more visible bet. The CEO presenting Plan A sounds more confident. The CEO presenting Plan B sounds, in the language most boards have been trained to hear, hedged.

But the two plans are not the same plan. They have wildly different distributions of outcomes. For a PE-backed company that needs to hit a specific exit multiple in thirty months, Plan B is almost certainly the right plan — the narrower distribution is worth more than the equivalent expected value, because the downside on Plan A blows the exit and the upside doesn't help enough to matter. For a venture-backed company chasing a power-law return, Plan A is the right plan — the fat tail is the entire point, and Plan B's narrower distribution is the wrong shape of bet entirely.

The board cannot have this conversation if neither deck talks about distributions. So the board has the wrong conversation, and the company makes the wrong choice — not because anyone made a bad call, but because the documents did not contain the information required to make a good one.

Two strategies with the same expected value can be radically different bets. A board that cannot tell them apart is not, in any meaningful sense, governing strategy. It is approving narratives.

What a bet-framed strategy looks like

The fix is mechanical. Take any strategy document and rewrite it, line by line, in the language of bets. Each major commitment becomes a row in a table with four columns: the bet, the odds we believe, the range of outcomes, and what would update us. The exercise takes a strategy team about a week the first time and about a day every quarter after.

For a real example — anonymized, but drawn from an engagement last year — the relevant rows of a rewritten strategy looked roughly like this:

Bet: Acquire the second-largest distributor in our category in Spain and integrate by Q4. Odds we believe: Sixty-five percent we hit the synergy case within eighteen months. Thirty percent we hit half the synergy case. Five percent we lose more in integration friction than we gain. Range: $40M to $110M of run-rate EBITDA contribution by year three, with a most-likely band of $70-85M. What would update us: If we have not closed the top three customer-retention commitments within ninety days of close, drop the central estimate by twenty-five percent. If the operations integration lead has not been hired by Q1, drop it by another fifteen.

Compare this row to its declarative equivalent in the original deck: Acquire Spanish distributor and integrate by Q4; expected EBITDA contribution $85M by Year 3.

The difference is not stylistic. The first version is a working document. A board can interrogate it. A CFO can build a cash-flow model around it that reflects the actual range of futures the company is preparing for. The CEO, six months in, has a pre-committed mechanism for revising the number that does not require an emotional re-litigation of the original decision. The second version is a press release that has not yet been issued.

Hamilton Helmer's 7 Powers makes a related point about durable strategic advantage: power is the persistent ability to earn returns above the cost of capital. The framing assumes you can tell, in the moment, which of your bets is plausibly building a power and which is simply riding a favorable market. You can't tell unless you have written down what you believed at the time, in language specific enough to be wrong.

Marc Andreessen, in his (less restrained) moments, has said something similar about technology bets: they are real options, priced by the willingness to write down the assumption that would have to hold for the bet to pay. Most strategic plans are real-options portfolios. Almost none of them are managed that way, because almost none of them are written that way.

What the board's role becomes

A board reading a bet-framed strategy has a different job than a board reading a declarative one.

The declarative board's job is to approve or reject. The deck either persuades or it doesn't. The discussion ranges over the deck's polish — the chart axes, the assumption sources, the competitive map. The board adds incremental rigor at the margin, then votes.

The bet-framed board's job is to govern a portfolio. The discussion is about which bets are sized correctly, which have early-update mechanisms in place, which are correlated with each other in ways the team has under-counted, which would benefit from being broken into smaller bets with explicit gates. The deck does not have to persuade; it has to expose. The board's value-add shifts from approving the narrative to pressure-testing the portfolio.

This is the role Kahneman and Lovallo, in their work on reference-class forecasting, argued senior leaders ought to play but rarely do. They observed that the inside view — built from the details of the case at hand — almost always over-estimates probability of success. The outside view — built from comparable cases at comparable scale — corrects this systematically. A board's structural advantage over the team that built the deck is, or should be, the outside view. But the outside view cannot be applied to a declarative deck. It can only be applied to a stated set of probabilistic claims. Eighty percent of acquisitions of this size in this industry hit half their synergy case within two years. Why does your case say sixty-five percent? What is different here? That is a useful question. It cannot be asked of the original deck, because the original deck never produced a number to interrogate.

This is why we treat the conversion to bet-framing as the highest-leverage change a strategy function can make. The team does not learn new analysis. It learns to state what it already believes in a form that the board can govern. The first cycle is awkward. The second is productive. By the fourth, it becomes difficult to imagine running strategy any other way — because the original way, once you have seen the alternative, looks like an organized performance of certainty rather than an act of leadership under uncertainty.

Where this fails

A few honest notes on where the bet-framing approach has limits.

It does not work on strategies that are genuinely vision-led, where the central claim is that a market will exist in five years and the team is being asked to bet on its existence. For these, the relevant numbers are unknowable in any meaningful sense, and dressing them in probability ranges produces false precision. The right document for that decision is a different one — closer to a memo defending the existence of the market and the team's privileged perspective on it. We treat those separately.

It does not work in cultures that punish written probabilities. If the CEO has built a career on declarative confidence and the board has rewarded it, introducing ranges reads as institutional weakness, and the first executive to write them down pays a career cost. In those settings, the change has to start at the board, not at the strategy team.

It does not, by itself, make a bad strategy a good one. A bet-framed bad strategy is a clearer view of a bad strategy. This is, in fact, the point — but it is worth saying, because we have occasionally been asked to install the discipline as if it were a magic that would make a doomed plan succeed. It will not. It will make the doom visible earlier, which is the next best thing.

The work

The work of converting a strategy from a forecast into a portfolio of bets is, again, mechanical. Most strategy teams can do it themselves once they have seen one done. We help when the cultural lift is too heavy for the team to do it alone — when the CEO needs to be persuaded that the new language strengthens, rather than weakens, the position; when the board needs a worked example of what the new conversation looks like; when the existing deck has too much organizational weight behind it to be honestly re-framed by the people who wrote it.

If the next strategic plan in your calendar reads as a forecast and you suspect, on inspection, that it ought to read as a set of bets, that is the conversation we have. The change in language is small. The change in what the company can see about its own future is not.