What to Read Next
Strategy is hard — really hard — to do well. Many leaders will admit this privately: In an anonymous 2019 survey conducted by Strategy&, 37% of 6,000 executive respondents said that their company had a well-defined strategy, and 35% believed that their company’s strategy would lead to success.1
Great leaders create ways of engaging their teams that can cut through this strategic fog. They may adopt frameworks to guide their analysis, but they expect participants in strategy discussions to contribute coherent reasoning and defensible ideas. Amazon is well known for its requirement that major initiatives be proposed in the form of a six-page memo. The virtue of the memo — versus a slide deck — is that writing in full sentences and paragraphs forces leaders to clarify how their ideas connect to each other. Similarly, Netflix has driven stunning transformations in the media landscape in part through its success at encouraging its leaders to debate ideas frankly and its willingness to empower them to take risks without waiting for an annual strategy planning process. It is no surprise that CEO Reed Hastings views working from home as “a pure negative” for the company, in part because “debating ideas is harder now.”2
Research Updates From MIT SMR
Get weekly updates on how global companies are managing in a changing world.
Please enter a valid email address
Thank you for signing up
The emphasis on vigorous debate at Netflix and Amazon clarifies a truth that many approaches to strategy obscure: At their core, all great strategies are arguments. Sure, companies can and do get lucky; sellers of hand sanitizer, for instance, have done very well during the pandemic. But sustainable success happens only for a set of logically interconnected reasons — that is, because there is a coherent logic underlying how a company’s resources and activities consistently enable it to create and capture value. The role of leaders is to formulate, discover, and revise the logic of success, making what we call strategy arguments.
Many leaders would agree with this claim but struggle with how to translate the insight into practice. What does it mean to construct a strategy argument? How does one evaluate such an argument?
In helping executives answer these questions, we have developed a flexible system of three activities: constructive debate, iterative visualization, and logical formalization. This system is facilitated by a set of concrete activities that can be used to develop a great strategy and execute it on a day-to-day basis. These practices demystify strategy and empower leaders throughout the organization to take control of their strategic destiny. They are neither mystical incantations nor rocket science; some of them have been practiced by logicians and philosophers for several millennia. But typically, they are omitted from the process of corporate strategy making.
The process of creating a great strategy is very different from how that strategy is communicated after its discovery. Put another way, executives should learn how to think like Jeff Bezos and Reed Hastings when tackling new problems and not simply study the decisions they made. There is no need to wait for a visionary savior to create great strategy.
The Ways and Means of Constructive Debate
Many leaders avoid arguing about strategy at all costs. Arguing is equated with fighting and, at best, is considered an unproductive use of people’s time.
This is a mistake. Arguing is the best way to do strategy, especially in groups, provided the arguments follow established rules of engagement that are rooted in the principles of deductive logic. Great strategy demands the exchange and vetting of ideas — both in its development and implementation.
Listen to Patty McCord, former chief talent officer at Netflix, who asserted, “The main reason the company could continually reinvent itself and thrive, despite so many truly daunting challenges coming at us so fast and furiously, was that we taught people to ask, ‘How do you know that’s true?’ Or my favorite variant, ‘Can you help me understand what leads you to believe that’s true?’” Such questions spawned vigorous internal debates at Netflix that, McCord said, “helped cultivate curiosity and respect and led to invaluable learning both within the team and among functions.”3
Why is debate so powerful? One reason lies in the fallibility of human reasoning. People typically argue inconsistently and are prone to confirmation bias — the tendency to interpret evidence in ways that confirm their preexisting beliefs. But while people suffer from confirmation bias when making their own arguments, it makes them better at evaluating the arguments of others: It causes them to cast a skeptical eye on arguments with which they disagree.
Furthermore, arguing constructively generates buy-in from the people involved. Being heard and having your ideas taken seriously leads to higher levels of commitment, both to specific strategic decisions and the organization as a whole.
Finally, reasoned debate is the only way to change how people think. Constructive debate helps participants understand how others in the group see the situation and allows them to arrive at a shared way of thinking. The result: superior strategic execution and learning.
To ensure that debates are constructive and arguing doesn’t devolve into browbeating and power plays, they need to be orchestrated. This requires thinking about strategic conversations before they begin and answering questions such as the following:
What is the purpose of this conversation? Devote time and attention, perhaps in consultation with others, to specifying the scope and desired outcome of the strategic dialogue in advance. For instance, if leaders are concerned about a new competitor, the team should debate whether the entrant truly poses a threat before considering costly initiatives to counter it. Accordingly, the purpose of the meeting would be to articulate the strategy argument of the new competitor and determine whether it interferes with the company’s own logic of success. Providing participants with clear guidance regarding the conversation (and assigning prework, if needed) helps them prepare for constructive debate and reduces their uncertainty about the parameters of the discussion.
Who should participate? Assemble people with relevant information and expertise, those responsible for executing the resulting decisions, and those with the power to make the decisions. Balance the need to represent different views and interests with the practicalities of conducting a productive discussion. We recommend including no more than 10 people, preferably fewer.
What are the participants’ roles? Most participants will arrive with ideas and proposals of their own, and the meeting should be structured to encourage everyone to share their views. We advise assigning just two roles in a strategic conversation: the facilitator and the devil’s advocate. The facilitator’s job is to get everyone involved by making the participants feel comfortable expressing their views. The devil’s advocate’s job is to challenge claims put forth and ask uncomfortable questions. Ask the remaining participants to leave their job titles and ranks outside the room before the start of the meeting. This is often a challenging ask for participants with the most formal power. Try reminding them that although the ultimate decision still rests with them, their authority should not be used to squelch debate. Instead, they should play subdued roles, not shoot down ideas, and allow their own arguments to be challenged by those with less power.
Where should the session take place? Choose a setting that offers the participants freedom of movement to encourage new ways of thinking. There should be enough space to rearrange the seating and multiple whiteboards for sharing ideas. If the meeting must be held online, keep the group small so that everyone remains engaged, and make use of online whiteboarding and collaboration tools.
How should the conversation be opened? Ask whomever decided to convene the discussion and the facilitator to choose a starter question to kick it off. Choosing a good starter question requires trying to imagine the different ways a conversation could go and how a question might be interpreted by the different participants. For instance, in a meeting regarding stagnant sales, a good starter question might involve asking each participant to arrive prepared to state what they believe the cause to be and why. Avoid yes or no questions, because they tend to polarize discussions and shut down a full consideration of the issues and processes at hand. Avoid specialized jargon as well, especially if it may not be understood by everyone in the room.
Strategy Maps Enable Iterative Visualization
Constructively debating a proposed strategy requires a shared understanding of the strategy. This means not only understanding the connections and causal flow between its elements, but also working creatively and collaboratively to harness different points of view.
During a July 2016 earnings call with Wall Street analysts, Southwest Airlines CEO Gary Kelly found himself battered by hostile questions. Analysts wanted to know why Southwest, which was experiencing revenue and profit declines, was sticking to its long-standing policy of first-come, first-served seating instead of following the lead of other airlines and charging passengers a fee for seat assignments. The analysts argued that Southwest was leaving money on the table. But Kelly dismissed the idea, claiming that nickel-and-diming passengers ultimately would not increase the airline’s profitability.
One might imagine that a similar debate had already taken place within Southwest’s headquarters as the executive team sought to reverse the airline’s declining results. In such a situation, the ability to argue constructively is critical to reaching the right decision. The intuition that seating fees would not help Southwest might be correct, but it should not be accepted simply because it feels right or because that’s what the boss believes.
Unfortunately, such debates devolve into shouting matches in many companies. Opponents hurl opposing justifications at each other with no clear means of resolution. Meetings drag on interminably as the opposing sides dig in. Finally, a decision is reached, but perhaps only because the most powerful person chooses the option he or she wants.
Instead, leaders can use strategy maps to surface and examine the underlying logic of their alternatives.4 A strategy map is a visual depiction of a strategy argument, using boxes and arrows to represent the structure of connections between ideas. It allows leaders to see how proposed investments and actions are linked to desired outcomes, how parts of a business model fit together, and how causes will supposedly lead to effects.
How might Kelly and his team at Southwest have used a strategy map to guide a discussion of seating fees? To start, put yourself in the team’s shoes. You know that Southwest’s ability to perform quick turnarounds of its planes is critical to its cost advantage. When a plane arrives at a departure gate, passengers are lined up and ready to go, and every employee at the gate and on the plane — flight attendants, pilots, baggage handlers, gate agents — helps them board and get ready for departure. This teamwork fosters flexibility and better labor utilization. Indeed, Southwest flies more available seat miles per worker than other major carriers, despite its shorter flight lengths.5
With this in mind, any consideration of fee-based seat assignments requires you to speculate about what will happen if Southwest abandons open seating: Will profits go up or will the impact be different? Such speculation involves visualizing scenarios — in this case, the current reality of open seating and a counterfactual one with assigned seating — to see how a change might reverberate through the system.
The boxes and arrows connected by solid lines in the figure “Mapping Out Strategy Arguments” depict how teamwork at Southwest’s gates improves capacity utilization and lowers staffing requirements. The boxes reflect ideas, concepts, resources, or actions, while the arrows represent causal claims. An arrow from A to B asserts that A causes B — that you believe the teamwork at the gate lowers costs through two distinct channels, one leading to greater capacity utilization and one leading to lower staffing levels.
Using the map allows the team to envision the impact of assigned seating. Moving away from open seating would increase the workload of gate agents. At other airlines, gate agents spend time responding to questions about seat assignments and addressing seat-change requests, upgrades, and other customer inquiries. Southwest gate agents, in contrast, spend no time worrying about who sits where. Accordingly, the map uses a dotted arrow to show the effect of open versus assigned seating on teamwork.
Switching from open to assigned seating might also affect passenger behavior. The boarding processes of other airlines can seem interminable as passengers board by section, search for their seats, and then jockey around getting settled. Boarding on Southwest is not painless, but first-come, first-served seating does speed things up. Hence, the dotted arrow is used to show the effect of open versus assigned seating on turnaround time.
It may seem just as easy to talk about alternatives as map them, but we’ve found that there are several benefits to strategy mapping.
First, strategy maps improve communication by clarifying complex relationships between the elements of an organization’s strategy. It is not obviously true that seat assignments would increase costs at Southwest. The analysts advocating the move felt that it should be easy to do. Creating a strategy map does not resolve the debate, but it does make clear how seat assignments may have negative operational consequences.
Second, strategy maps facilitate constructive debates. Seeing someone’s argument represented visually creates opportunities for people to recognize and express how their own views differ. An advocate of assigned seating would note that the strategy map ignores the increased revenue it would generate for Southwest, leading to a more productive exchange.
Third, strategy maps help surface unstated beliefs and assumptions. Creating a map does not guarantee success, but by clarifying assumptions, it does makes it more likely that a better decision about seating will be made. As important, if things do go wrong, being clear about the reasoning behind the decision makes it much easier to learn from mistakes and take corrective action.
A final benefit of strategy mapping is that executives typically find it an intuitive way of developing and representing their arguments. Indeed, many already sketch out their ideas in this way. Still, productive strategy mapping takes discipline and practice. Here are a few tips for using strategy maps:
Start with the conclusion, and map only one conclusion at a time. Start by stating clearly what it is you are trying to explain: the destination at the end of the map. In addition, because group discussion often surfaces many issues, it is important to clarify from the outset which conclusion will be the focus of the mapping exercise. Limiting the scope of the exercise helps avoid emotional and unpredictable disputes.
Identify and organize all plausible explanations. Adopt a brainstorming mindset to generate as many ideas as possible, including some (or many) that are “outside the box.” For instance, at Southwest, the participants could have been asked to list all the activities and factors that enable the company to offer air travel at lower costs than its competitors. If the number of ideas starts to seem unmanageable, cluster them into related themes and concepts. This is particularly easy to do if the initial ideas have been captured on sticky notes. The quality of a strategy map depends on the how comprehensively it captures the causes (conditions, assumptions, or premises) that might plausibly generate the map’s conclusion.
Construct the map. Combine and recombine the ideas on sticky notes to show how inputs lead to outputs or causes to effects. This iterative process enables the participants to explore the relationships between different ideas in a flexible, dynamic manner. The key to constructing a map is to focus on the facts and circumstances needed to generate a desired outcome so that the basic structure of the map, once completed, reflects the basic structure of the strategic argument.
Formalizing the Strategic Logic
Strategic debates within companies tend to rely heavily on informal exchanges. Executives express ideas about the effects of a new rival, a government policy, or an emerging technology on a strategic outcome and link it to a particular course of action. They are making arguments, but the arguments are usually incomplete. They lack internal coherence — what logicians call validity.
Great strategies exhibit logical coherence. They are composed of a set of logically interconnected reasons that necessarily produce the conclusion. With a valid argument, if you accept the premises, you must accept the conclusion. With an invalid argument, you run the risk of overlooking critical assumptions and flaws in your reasoning, which in turn can lead to your company’s downfall. Logical formalization ensures validity.
When Apple released the first iPhone in 2007, executives at Nokia, then the leader in the mobile phone market, thought the new product had little chance of success. Why? They believed that the iPhone was an inferior phone. Its call quality was poor, and calls were more likely to be dropped. Its use was limited to 2G networks, whereas the leading phones were using the 3G standard. Other cellphones had longer battery lives, were more rugged, and fit neatly into a pocket. No wonder Nokia’s chief strategist concluded that the iPhone would be a niche product.
The executives at Nokia — and many other experts at the time — were applying flawed logic. Imagine the debates that likely took place. “The iPhone does not have good cellular phone technology, so it will not win in the cellular phone market,” asserts the skeptic. “You’re wrong! Apple has an amazing design sensibility, so of course they will dominate the cell phone market,” argues the believer. “Come on! It’s a lousy phone; there is no way they can win,” responds the skeptic. And so on.
This kind of exchange, familiar as it may be, is rarely productive. Such debates lead to bad decisions because of flawed logic.
The problem starts with the initial claim (“The iPhone does not have good cellular phone technology, so it will not win in the cellular phone market”). Like many informal claims, it is an incomplete argument — critical assumptions are not stated. As a result, the argument is logically invalid. The conclusion (“… so it will not win in the cellular phone market”) does not necessarily follow from the premise on which it is based (“The iPhone does not have good cellular phone technology”).
Unfortunately, incomplete arguments of this type are not only common, they also can carry a lot of weight in strategic debates. Leaders and their audiences do not notice their gaps, because they appeal to implicit premises in their minds. Instantly and subconsciously, their minds bridge the gap with these implicit premises. This makes for easy conversation and lively debate, but it undermines good reasoning.
Strategy mapping encourages leaders to be more explicit about their claims. But maps can also mislead by leaving out critical elements or failing to represent critical facts about the organization or its environment. This is where logical formalization comes in.
Assessing validity is easier when strategy arguments are stated in propositional form. If you do this with the initial claim in the iPhone example, you get two statements:
Statement 1: The iPhone does not have good cellular phone technology.
Statement 2: The iPhone will not win in the cellular phone market.
The first statement is a premise — an empirical assumption about the iPhone that may or may not be true — and the second is the conclusion. If Statement 1 is true, then Statement 2 is true. The logical problem here is that nothing connects the two statements. If you don’t see this, try substituting a different empirical claim for Statement 1, such as “The iPhone was not designed in Finland.” Suddenly the gap is obvious.
Logical formalization involves making the argument complete by explicitly stating how you believe Statements 1 and 2 are connected to each other, as so:
Premise A: The iPhone does not have good cellular phone technology.
Premise B: If a product has good cellular phone technology, then it will win in the cellular phone market.
Conclusion C: The iPhone will not win in the cellular phone market.
The claim is now complete (there are no missing premises), but the argument isn’t logically valid yet. It is easy to see why the conclusion does not necessarily follow: We know of many examples where the technically superior product has not won in the market. It could be that none of the products in the market has good technology, and yet there will be a winner.6
Still, the restated argument offers progress, because now we know what to fix. A logically valid argument could be constructed by revising Premise B:
Premise B*: If a product does not have good cellular phone technology, then it cannot win in the cellular phone market.
The revised argument clarifies exactly what must be true to accept the original claim: that the iPhone will fail because it has poor cellular phone technology.
The widespread agreement that the first iPhone would not succeed because it was not a very good phone was mistaken.7 At the time, however, the truth of Premise B* was unknown — reasonable people could disagree. Indeed, this is likely one point where Apple and Nokia executives differed. Given that Apple released a phone with inferior phone service, its executives probably did not think that dropped calls precluded success. Their theory of success likely rested on a different argument — a competing theory of what it would take to win. Nokia executives, on the other hand, could reasonably believe that good cellular phone technology was a prerequisite for success. But they and executives at other cellular makers would have been wise to examine their assumptions and consider why Apple’s executives’ beliefs differed.
Surfacing implicit assumptions and ensuring validity will not tell you what the future will hold. But by using logical formalization to identify what must be true for a certain vision of the future to hold, you can shift the terms of debate in productive ways. It helps bring the critical issues into focus, and it will help members of your strategy team move from fighting about whether they believe a conclusion to recognizing what they have to believe in order for it to be true.
Formalizing an argument is often challenging and may seem awkward. Here are two tips that can help:
Start with the easy stuff. As in the iPhone example, begin by trying to fill in the missing premises in an informal claim. If the informal claim is “A is true, so B will happen,” convert it into an if-then statement (“If A, then B”) to make the causality clear. This may seem trivial, but an if-then statement represents a theory of how the world works, and stating it explicitly forces people to consider whether it holds under all circumstances.
Be prepared to iterate. Keep in mind that the goal in formalizing a strategy argument is not the unassailable perfection that a logician seeks, but a deep, shared understanding of what your colleagues are saying. Often, people will have different ways of phrasing new linking premises; use those differences to illuminate conflicting ideas and arrive at a more comprehensive and collective understanding of the argument.
Strategy requires making major decisions with lots of moving parts and often far-reaching and hard-to-predict consequences. The decisions entail trade-offs: having the courage and the discipline to say “no” to one course of action in the uncertain hope that another will be more promising. Frequently, these choices are irreversible — and their outcomes are likely to affect the careers of the leaders making them. Yet, as we have seen, a majority of leaders lack confidence in the soundness of their companies’ strategies.
Much of the problem with strategy-making arises because executives focus on predicting what the world will look like in the future when debating strategic decisions. This is unproductive and frustrating; even the most accurate predictions are worth little if they are not embedded in a logically coherent strategy argument. Apple made the right bet about the relative importance of call quality with the original iPhone, but the bet paid off only because it was part of a valid argument that recognized what else had to be true for the iPhone to win.
The core of every great strategy is a valid argument. Leaders should work to develop such an argument. It should include what needs to be true for the strategy to succeed, and it should specify the resources (people, technology, finances, and even regulators) required to execute it. Constructive debate, iterative visualization, and logical formalization are essential tools for arriving at such arguments. They are a practical and accessible way for working executives and their teams to develop the skills and habits of mind needed to achieve strategic success.
1. “The Strategy Crisis: Insights From the Strategy Profiler,” PDF file (New York: Strategy&, 2019), www.strategyand.pwc.com.
2. J. Flint, “Netflix’s Reed Hastings Deems Remote Work ‘a Pure Negative,’” The Wall Street Journal, Sept. 7, 2020, www.wsj.com.
3. P. McCord, “Powerful: Building a Culture of Freedom and Responsibility” (Jackson, Tennessee: Silicon Guild, 2017), 52-53.
4. Our use of the term “strategy map” should not be confused with that of Robert S. Kaplan and David P. Norton in “Strategy Maps: Converting Intangible Assets Into Tangible Outcomes” (Boston: Harvard Business School Press, 2004). They use the term to refer to a diagram based on a balanced scorecard approach to strategy.
5. “Airline Industry Overview,” Global Airline Industry Program, Massachusetts Institute of Technology, accessed April 12, 2021, http://web.mit.edu.
6. Logicians call this “denying the antecedent,” and it takes this form: Premise 1: If X, then Y. Premise 2: Not X. Conclusion: Not Y.
7. S. Gilbertson, “IPhone First Impressions: Not Worth the Money,” Wired, June 29, 2007, www.wired.com.