The Power of the Strategic Initiative Postmortem

Jackson Pollock, Alchemy (1947), Peggy Guggenheim Collection, Venice

Contents

• Introduction
• Why Strategic Initiative Postmortems Are Rare
• Three Common Failure Modes
• What Makes an Effective Strategic Initiative Postmortem
• Shopify 2022
• The CEO Sets the Bar
• Guidance for Strategy Leaders
• Worksheet

Introduction

In my last post, I discussed Frank Keil’s work on causal understanding and its relevance to strategy. With this one, I want to give you a very practical way of building causal understanding, at a particularly valuable moment: after a strategic bet has run its course.

Strategic initiatives end all the time. Some succeed, some fail, and many quietly die. In Chapter 26 of our forthcoming book Strategy and Value (out November 17), just published in McKinsey Quarterly in slightly compressed form, we say that joining strategic initiative postmortems is one of eight practices we recommend to CEOs choosing how they engage in strategy. The purpose is not to assign blame or shame. It is to hold a high bar for learning. This focus on learning from the past distinguishes many great CEOs. A. G. Lafley, while leading Procter & Gamble, went as far as having 30 years of unsuccessful acquisitions analyzed to understand their root causes.

In my experience, strategic initiative postmortems rarely happen. When they do, they usually don’t happen at a strategic level, instead focusing on simple explanations of what went wrong at an operational level. A strategic initiative postmortem asks three harder questions:

1. What did we assume about the future when we committed, and why?
2. What actually happened, and why?
3. Which of our beliefs does this challenge or reinforce, and why?

This discipline makes compound learning in strategy possible. With many leadership teams preparing their 2027 calendars, now is a good moment to ask whether they are doing enough to learn from the past.

Why Strategic Initiative Postmortems Are Rare

There are many reasons why they are rare.

Strategy is reviewed in cycles, not as each choice merits it. Many companies revisit their strategy as a whole each year, rerunning a standard process to (re)produce an integrated strategy. In such instances, a postmortem can feel burdensome, like an extra process step between two years’ strategy processes. Even worse, if the strategy doesn’t have clearly defined choices and is more thematic, then the idea of running a postmortem on part of it might feel intimidating, if not impossible.

Companies that view their strategy as a portfolio of strategic initiatives, as we describe in Strategy and Value, find it more natural to run strategic initiative postmortems, because each strategic choice stands on its own with its own timeline. When a strategic initiative is retired, the postmortem is due, whenever that may be.

There is no clear record of what was assumed when a choice was made. When the key assumptions behind a choice were never written down, a strategic initiative postmortem must first reconstruct them, from scattered documents and from memories that hindsight has already begun to edit. That is a slow, inefficient, and often contested process that begs to be set aside for more urgent work.

Strategy Champions tend to be more efficient because they are more likely to document those assumptions when they commit. As a result, preparation for a strategic initiative postmortem can focus on comparing what was assumed with what happened, rather than on reconstructing the past.

Looking back feels like time taken from the future. Slowing down to unpack what happened can feel significantly less useful than focusing on the hard, uncertain strategic questions of the future. This is particularly problematic in companies that don’t have a strong learning-oriented culture.

The social side of strategy resists it. The people who owned or championed a strategic initiative may have moved on and have no real incentive to revisit the topic.

The state of debate doesn’t support it. Just because the past is behind us doesn’t mean uncertainty is eliminated. A strategic initiative postmortem requires a team to effectively debate what happened and why, and while there may be more evidence to support the dialogue, that doesn’t make it easy. Organizations that are either overly consensus-oriented or don’t provide sufficient psychological safety will struggle to do this. AI, when used incorrectly, can make this worse, as a new HBR article on the effects of employees turning to AI before teammates points out.

Three Common Failure Modes

When strategic initiative postmortems do happen, three failure modes are the most common.

The strategy is blamed, not its quality. “The strategy was flawed” is an easy conclusion. It is also important to ask whether the strategy as a whole and the individual strategic initiative were held to a high enough bar. Was there a clear value creation thesis or theory that helped people understand the logic? Was the choice made and codified clearly enough to be effectively owned by a single leader? Was the choice bold enough, and did it name what would stop? A strategy can contain a kernel that was right and still fail, because it was never defined well enough to have a chance of being mobilized.

Mobilization goes unexamined. Failure gets attributed to execution: missed timelines, the wrong pace, poor coordination, barriers not being broken. What is less often asked explicitly is whether the strategic choice was ever truly mobilized. Was there a quality handshake to the single leader that would own it, and were their targets and incentives updated quickly? Were all the necessary resources actually reallocated, quickly and sufficiently? Did budgets and plans get updated to reflect its priority, and were things actually stopped to support it? As we write in Strategy and Value, what companies most often diagnose as an execution failure is a mobilization failure.

The story is oversimplified. Postmortems drift toward tidy explanations with a single cause. Research by Saku Mantere and colleagues found that the different people involved in a failed venture each tend to construct their own narrative of what happened, often one that serves self-justification. Hindsight makes this worse. Once we know an outcome, we overestimate how predictable it was and misremember what we believed beforehand. In a recent New York Times op-ed calling for business schools to be abolished, Christopher Marquis, who spent a decade writing cases at Harvard Business School, made a related critique: some cases reflect views of the past that are a little too tidy.

What Makes an Effective Strategic Initiative Postmortem

A strategic initiative postmortem asks the three questions above, and treats the strategic choice itself, not only its execution, as the subject of review.

Assumed. Reconstruct what the team assumed about the future when it committed, from the record rather than from memory. This is why documenting the key assumptions behind each strategic choice matters so much. In our Strategy Champions research, it was the sixth-largest differentiator among 28 capabilities, with top performers 25 percentage points stronger than stragglers. As mentioned, the record unlocks efficiency and fights hindsight bias.

Learned. Diagnose what happened across five possible sources:

1. The design. Which assumptions broke, and across which of the four foundations we describe in the book: market attractiveness, competitive advantage, growth and ROIC, or managing for the long term? Trace the chain of hypotheses behind the strategic choice, link by link, to find where the logic failed.
2. The quality of the design. Was the strategic choice held to a high enough bar before it was committed to? A sound idea, poorly defined, can fail without ever being properly tested.
3. Mobilization. Was the strategic choice effectively mobilized? Check whether one leader owned it with the right targets and incentives, and whether resources actually moved.
4. Execution. Were the supporting plans, projects, and efforts carried out well, were barriers broken and focus sustained, and did learning and adaptation happen at the right times? Examine it last rather than first, because it is where blame lands most easily.
5. Luck. What lay outside anyone’s control? A good bet can lose and a poor one can win, so judge the decision separately from the outcome, avoiding what Annie Duke calls “resulting.” And don’t let bad luck prevent learning.

Successes deserve the same scrutiny. In a study with Israeli soldiers, Shmuel Ellis and Inbar Davidi found that those who reviewed both their successful and failed exercises improved more than those who reviewed only their failures.

Updated. The third question shifts from the assumptions behind one choice to the beliefs behind many choices. Assumptions are judgments about a single situation; beliefs are broader views about how strategy works. One failed bet rarely overturns a belief, and one success rarely proves it, so ask how strongly the evidence challenges or reinforces each one. As we describe at the end of our excerpt from Strategy and Value, this is how the best CEOs build their beliefs (and values): by making decisions, watching them play out, and taking time to properly reflect on what they learned.

Though you should push to get to the few reflections that matter most, don’t strip out the discomfort, complexity, and chaos of reality. Messiness is what helps people build pattern recognition about what they will likely face in the future.

Shopify 2022

Public postmortems are a unique case, because they are also communications to employees and investors. They can be powerful teaching tools when done right.

Get Whitney Zimmerman’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

One exceptional (if unfortunate) example is the memo Shopify’s founder and CEO Tobi Lütke sent to employees in July 2022, when the company let go of about 10 percent of its workforce.

Lütke began with what Shopify had believed. During the pandemic, the company had bet that ecommerce’s share of retail would permanently leap ahead by five or even ten years. He reasoned that if there was even a chance this was true, Shopify would need to expand to match, and it did.

He was equally candid about what Shopify had learned. The bet had not paid off. The mix was reverting to roughly where pre-pandemic trends suggested it would be.

One way to read his memo is through our four foundations of strategy and value, each of which it hits:

• Market attractiveness: The belief concerned growth quality, meaning whether the surge in ecommerce reflected a structural shift or a temporary one. It proved temporary.
• Competitive advantage: He explained why the bet mattered so much. Shopify’s market share in ecommerce is much higher than in retail, so the question was how much of the market would move to where its position was strongest.
• Growth and ROIC: The company had built its cost base for growth that did not arrive. The memo connects the broken belief directly to the resource decision it justified.
• Managing for the long term: He noted that ecommerce was still growing steadily. The error was about pace and permanence, not direction, so the long-term thesis could stand while the overreach was corrected.

His memo is particularly powerful because it names specific, measurable assumptions, clarifies the errors, connects them to the decisions they justified, and takes personal ownership of them with candor and without blaming execution. And it does all of this succinctly.

Unsurprisingly, Shopify’s engineering teams have also been known to run what are called “Assumption Slams” to surface the assumptions behind major programs at the outset, as I noted in my post on compound learning. A company that has a habit of naming its key assumptions up front is better placed to name the ones that broke.

Because Lütke framed the expansion as a bet under uncertainty, it also opens a deeper question for the room. Was it wrong to make the bet, or right to make it and wrong in how much was staked? That distinction between the decision and the outcome, worked through with the messiness left in, is exactly what a strategic initiative postmortem inside a company should tackle.

The CEO Sets the Bar

As mentioned, our new perspective on the CEO’s six undelegatable roles in strategy recommends that CEOs (or the equivalent for the relevant part of the organization in question) insist on strategic initiative postmortems and participate in them. This is because the CEO’s involvement is uniquely valuable. It also reflects a point made by A.G. Lafley in his perspective on what only the CEO can do, in being responsible for setting standards and norms in the organization. Learning from strategic choices is an important norm that demands a standard of excellence.

Inside the room, the CEO sets the standard of explanation by asking why as often as needed, and pushing past the first comfortable answer. It means asking explicitly about mobilization and the quality of the strategy, and refusing tidy single causes. The CEO will bring and model their own beliefs, of course, but what they ask of others matters more.

Outside the room, how a CEO talks about strategic bets in town halls, memos, and letters shapes organizational norms. Lütke’s memo is exactly this kind of outside-the-room document.

The right stance is blameless toward people but rigorous toward beliefs and assumptions. Engineering teams, following practices like Google’s blameless postmortems, assume everyone acted reasonably with the information they had. That is the right stance toward people, and it is essential for psychological safety and effective debate.

What This Means for Strategy Leaders

1. Manage strategy as a portfolio of strategic initiatives, so that each one has a natural end point that prompts a postmortem, of successes as well as failures. If you look at your strategy and find it difficult to determine when a specific choice would merit a postmortem, then you should question whether your strategy meets the bar as defined. Revolut founder and CEO Nik Storonsky talks about a similar approach to bet-based strategy in his latest interview with David Senra.
2. Document key assumptions when you commit to a strategic choice.
3. Diagnose mobilization and the quality of the strategy explicitly. Execution is rarely the whole story.
4. Embrace the messy version, and where appropriate, turn it into an internal case that teaches your approach to strategy.
5. Feed what you learned into the next strategy, starting with a premortem. Before committing to the next bet, imagine it has already failed (or succeeded) and ask why. That is how learning truly compounds.

And in general, I recommend reflecting on when the next opportunity to run a postmortem on a strategic initiative will be, and whether your 2027 approach to strategy sets you up for successful learning.

Worksheet

Readers who preorder Strategy and Value will get access to a set of extra worksheets, and writing this post prompted me to add one more: a worksheet on running a strategic initiative postmortem. I will share the official announcement and instructions soon. If that sounds useful, you can go ahead and preorder the book here.

A Note on the Artwork

Jackson Pollock, Alchemy (1947), Peggy Guggenheim Collection, Venice

I like to contrast Mark Rothko with Jackson Pollock when talking about what a clearly designed strategy looks like. Rothko, whom I wrote about in an earlier post on why strategy is about choices, was known for color field paintings that contain a few large, clearly bounded fields of color. Like a good choice-based strategy, a color field painting has clear choices with well-defined boundaries (reflecting what will not be done), they work together with an underlying theory (as Rothko was a color theorist and thought deeply about how colors interact to impact the viewer), and they have a profound impact on viewers. Too many strategies look more like a Pollock: themes such as digital, growth, and productivity dripped across the whole canvas.

Clear choices, like Rothko’s color fields, are much easier to isolate for review and revisit when they run their course. When a strategic initiative ends, you can go straight to the choice and key assumptions and context behind it. When a strategy is a web of themes or many, many small decisions, as in a Pollock, there is no single line to follow.

I don’t mean this as a criticism of Pollock. He was adamant that his work wasn’t chaos. In 1950, Time quoted the Italian critic Bruno Alfieri’s verdict on the Pollocks shown in Venice that summer, Alchemy likely among them, under the headline “Chaos, Damn It!” Pollock cabled back: “No chaos damn it.” Alchemy is built from dense layers of paint, with twine, sand, and pebbles worked into the surface. When conservators studied and restored it, they found Pollock had followed a rational plan in laying down its nineteen pigments. The order is there, but it takes patient, layer-by-layer work to see it.

That makes a Pollock a good picture of the past itself. Even when the choices were clear, what actually happened rarely is. Causes overlap, and early layers are partly hidden by later ones. Clear choices don’t make that history simple, but they tell you where to start looking. Design your strategy to look more like a Rothko; don’t be surprised if its history looks more like a Pollock; and resist the urge to repaint that history in tidy blocks as you invest time and care to learn from it.

To Go Deeper

• Andy West, Tim Koller, Whitney Zimmerman, et al., “The six strategy roles a CEO must not delegate,” McKinsey Quarterly
• Tobi Lütke, “Changes to Shopify’s team,” Shopify, July 26, 2022
• Shmuel Ellis and Inbar Davidi, “After-Event Reviews: Drawing Lessons from Successful and Failed Experience,” Journal of Applied Psychology 90, no. 5 (2005)
• Saku Mantere, Pekka Aula, Henri Schildt, and Eero Vaara, “Narrative Attributions of Entrepreneurial Failure,” Journal of Business Venturing 28, no. 4 (2013)
• Google, “Postmortem Culture: Learning from Failure,” Site Reliability Engineering
• Kate Niederhoffer, Kristina Rapuano, Alex Liebscher, Alexi Robichaux, and Jeffrey T. Hancock, “Team Dynamics Weaken When Employees Turn to AI Instead of Colleagues,” Harvard Business Review (October 2026)

My previous posts referenced above

• Holding on to the ‘Why’ in Strategy: A Conversation with Yale Cognitive Scientist Frank Keil (LinkedIn, Medium, Substack)
• Strategy Must Survive to Matter: Lessons from Andrew Marshall, the Pentagon’s “Yoda” (LinkedIn, Medium, Substack)
• New Book Preview: Strategy and Value (LinkedIn, Medium, Substack)
• The CEO’s undelegatable roles in strategy — a Strategy and Value excerpt in McKinsey Quarterly (LinkedIn, Medium, Substack)
• Unlocking Compound Learning in Strategy (LinkedIn, Medium, Substack)
• Strategy is About Choices (LinkedIn, Medium, Substack)