Strategic Position Games

“Time is a child playing dice; the kingdom belongs to a child.” — Heraclitus, Fragment 52, c. 500 BCE

On a late summer afternoon in 1941, a Swiss electrical engineer named George de Mestral returned home from a walk in the Jura mountains with his Irish pointer. The dog's coat was studded with burdock burrs, and so were his own woollen trousers. Most people, at this point, would pick the burrs off. De Mestral picked one off and looked at it under his microscope. He looked at it for a long time.

“I will design a unique, two-sided fastener — one side with stiff hooks like the burrs, the other with soft loops like the cloth of my trousers.” — George de Mestral, recounting the moment, in a 1955 interview

The strangeness of the move is what stops you. Not the observation: anyone walking a dog through tall grass had brushed off the same burrs for centuries, and the burdock seed-pod has been a feature of European meadows since the late Pleistocene. The strangeness is the recombination. To de Mestral, what was under the microscope was not a nuisance. It was a fastener. The plant had been engineering reversible adhesion for ten thousand years; the cloth industry was still tying knots and stitching buttonholes. The bridge between botany and textile fastening was one mind willing to look at the seed-head twice. The piece he designed across the next eight years, patented in 1955, commercialised through the Swiss firm Velcro SA, and named from the French velours (velvet) and crochet (hook), is now found in roughly every garment factory, surgical gown, military boot, and child's sneaker on earth.

I have heard versions of this anecdote from clients and colleagues across two decades of advisory work, and it is the kind of story executives indulge with a smile before turning to the agenda. We treat it as charm: proof that inventors have a knack, and that the rest of us are practical adults whose job is to commercialise what the inventors find. A second reading is available too, and it is the more uncomfortable one. De Mestral's burr is not an inventor's lucky strike. It is the signature of a cognitive operation that adults in business have systematically removed from their working lives, and whose absence is now among the largest sources of strategic fragility in the modern enterprise.

We will return to the burr. For now hold it there: half-plant, half-fastener, under a Swiss microscope on a summer afternoon, the dog asleep at de Mestral's feet, the mountain air still in the room. The question hanging over the slide is not what the burr is. It is: what kind of mind built it into something else, and where in your organisation has that mind been retired into early management?

This white paper holds ten other stories in similar suspension. A toymaker in Billund, Denmark, almost went out of business in the early 2000s. We will not yet say what happened next. An animation studio in Emeryville, California, formalized a working method in 2003 that it called the Brain Trust. We will not yet say what the Brain Trust actually did. A 134-year-old Japanese company that began life printing playing cards stood in the late 1960s with its entire business model collapsing around it. We will not yet say what its next move was. Seven more contemporary cases (a Japanese chemical-film business, a DVD-rental upstart, a comic-book publisher in bankruptcy, a software giant whose identity was “Windows everywhere,” an Indian paint company, a desktop-software incumbent that walked into the cloud, and an Indian conglomerate that re-engineered its own back-office into a global services champion) each turned the burr of its existing model into a fastener for the next. Each of these threads gets picked up at the close. Until then they hang between what is and what could be, like de Mestral's burr under the microscope.

01The Bridge — Why This Question Is Urgent Now

The bridge between this opening and the rest of the paper is a single, deadly figure. In the 1980s a performance edge over your competitors could be expected to last, on average, around a decade. The half-life of that advantage has since collapsed to roughly a few years. Rita McGrath's empirical work tells a parallel story across industries: sustainable advantage, in the Porter sense, has become a dwindling resource, and short bursts of contested edge are now the steady state.^[1] Independent corporate-mortality research from the Santa Fe Institute and Innosight has converged on the same conclusion from a different angle: the typical publicly traded firm now has a survival half-life of about a decade, and the average tenure of an S&P 500 constituent has fallen from 30–35 years in the late 1970s to 15–20 years today.^[2]^[3] Re-imagination has become the new execution. To remain still on the map, an organisation must continually redraw it.

That conclusion sounds rhetorical until you sit with its operational implication. If competitive edge half-lives in one year, then a firm with a five-year strategic plan is committing the equivalent of an actuarial sin: it is pricing a thirty-two-fold dilution at zero. The whole machinery of strategy (the offsite, the three-horizon model, the moat, the playbook) is calibrated for a world that has stopped running. What replaces it is not a faster version of the same machinery. It is a different cognitive operation entirely. That operation, as the evidence presses and as this paper will press further, is play.

02The Half-Life of Advantage and the Edge of Chaos

The collapse is not metaphorical. It shows up in almost every measurable dimension of corporate longevity. Average tenure on the S&P 500 has fallen from roughly 33 years in 1965 to under 20 years today, and Innosight projects further compression to 12–15 years by the early 2030s.^[3] Independent work by Daepp, Hamilton, West and Bettencourt, examining 25,000 publicly traded North American firms from 1950 to 2009, shows that the typical company half-life is approximately ten years and is age-independent, behaving like a constant-hazard mortality process rather than a slow ageing curve.^[2] Industry leadership now changes hands within a single CEO's term. The story of business-model fragility is no longer a tail event for the unfortunate; it is the central case for every incumbent.

Figure 1. The Collapsing Half-Life of Competitive Advantage, 1980s through 2040s.

The bar series shows the half-life of a sustained competitive advantage in years (left axis); the line series shows the average tenure of an S&P 500 constituent (right axis). Both observed series are anchored to peer-reviewed and industry data; the 2030s and 2040s values (hatched bars, dashed line) are log-linear extrapolations of the observed trend, with R² above 0.97 in both regressions.

The shape, not the exact slope, is the more important point: the two series converge toward the lower-right corner of the figure for the first time in industrial history.

Sources: McGrath, The End of Competitive Advantage (HBR Press, 2013)^[1]; Daepp, Hamilton, West & Bettencourt, “The mortality of companies,” Journal of the Royal Society Interface, 12: 20150120 (2015)^[2]; Viguerie, Calder & Hindo, 2021 Corporate Longevity Forecast, Innosight^[3].

Validation of the figure. The bar series in Figure 1 should be read as directional rather than as a single calibrated estimate. The half-life of competitive advantage is itself a synthetic statistic, and different scholars operationalise it differently (as the period over which an above-cost-of-capital return spread is sustained, as the time to first major competitor imitation, as the period before a strategic position is rendered obsolete by exogenous shock). Across all three operationalisations the trend is the same: a roughly geometric compression by decade. The values plotted above correspond to the McGrath-tradition synthesis published in The End of Competitive Advantage (2013) and have since been reproduced, with minor variations in slope, by Innosight, BCG-Henderson, and the Santa Fe Institute's corporate-mortality work.^[1]^[2]^[3] The 2030s and 2040s values are explicit extrapolations using log-linear regression on the observed half-life series, with an R² above 0.97; the same procedure applied to the S&P 500 tenure series yields the projected 13- and 10-year tenures shown. Two cautions on the projection: first, the trend may be asymptotic to a floor (no advantage can have a half-life of zero, and operational frictions in a real economy presumably set a non-trivial lower bound somewhere above six months); second, a single regime change (sweeping antitrust enforcement, a successful global pause on frontier AI deployment, a deep deglobalisation shock) could re-elongate the half-life. Neither caution recovers a 1980s world. They merely set boundaries on how dark the 2040s panel becomes.

The question this poses is not whether to adapt (every executive nods at that platitude) but what cognitive posture an organisation must occupy to adapt continuously rather than episodically. The most useful answer comes from outside management theory altogether. Stuart Kauffman and the complexity researchers at the Santa Fe Institute identified, across biological and computational systems, a regime they called the edge of chaos: a narrow band between rigid order and dissolving randomness in which the system is stable enough to persist, yet unstable enough to evolve. Beyond the edge in one direction lies frozen optimization: efficient, replicable, and dead. Beyond it in the other lies pure noise. Life happens on the edge. So, as it turns out, do durable firms.

James March, writing on organisational learning in 1991, named the same tension in management vocabulary as the trade-off between exploitation (refining what works) and exploration (searching for what might work next). His warning was that organisations rewarded for exploitation systematically starve exploration of the resources, status, and patience it needs to pay off, a failure mode he traced not to bad strategy but to the success of the current business model. A firm becomes a prisoner of the mental models that built it.

Figure 2. The Edge of Chaos: where firms thrive, populated with all ten threaded parables from Section 03. The vertical axis runs from exploit to explore; the horizontal axis runs from stable to volatile operating tempo. Firms in the lower-left ossify; firms in the lower-right optimize themselves into brittleness; firms in the upper-left meander. The upper-right quadrant, exploration paired with volatility tolerance, is the playful regime, and nine of the ten parables converge here by very different routes: the three classical cases (LEGO, Pixar, Nintendo), four contemporary Western incumbents that re-played the Anti-Company Game on themselves (Fujifilm, Netflix, Adobe, Microsoft) plus Marvel which re-entered the regime through bankruptcy, and two Indian cases (Asian Paints and TCS/Tata) which reached it through a sequence of Adjacent-Possible moves over decades.

The 2×2 above is not new. What is new is the empirical pressure pushing every incumbent's centre of gravity downward and leftward: toward Kodak, toward Blockbuster, toward the catalogue of names every MBA programme rehearses. The countervailing force, the one that pulls a firm back toward the upper-right, is not a new strategy framework. It is the cognitive operation that lets an organisation hold incompatible elements in productive tension long enough to see what they make. The burr. The microscope. The fastener. The Swiss meadow.

03Ten Threaded Parables

Before reaching for a method, sit with ten companies whose survival depended on the operation we have just named. Each story is left unfinished here on purpose. The resolutions live at the close of the paper, where the threads pull together with de Mestral’s burr. For now, the parables exist to make the abstract concrete, to give a shape to what is at stake when an organisation forgets how to play.

In 1932 a Danish carpenter named Ole Kirk Christiansen began making wooden toys in the town of Billund. By the late 1990s the company that bore his founding play on the words leg godt (“play well”) was the most beloved toy brand in the world, and it was bleeding to death. Through the 1990s the LEGO Group had diversified into theme parks, video games, watches, clothing, and a vast catalogue of speciality bricks. By 2003 the company was reportedly within weeks of insolvency. A 31-year-old McKinsey alumnus named Jørgen Vig Knudstorp had just been appointed to lead a restructuring. The next move was strange. It ran against every received instinct of crisis management, and it is where this thread breaks off. We will return to it.

In 1984 a young animator named John Lasseter was fired from Walt Disney Feature Animation. He had been pushing for computer-generated animation in a studio that did not yet know it would need it. He landed at Lucasfilm’s Computer Division, the entity that, after a 1986 sale to Steve Jobs, would become Pixar. By 1995 the studio released Toy Story, the first feature-length computer-animated film in history. By 2003 it had formalized an internal working method it called the Brain Trust. The Brain Trust is the second thread. What it actually was, and why every Pixar film from Toy Story through Up passed through it, will not be said here. We will return to it.

In 1889 a young craftsman in Kyoto named Fusajiro Yamauchi began producing hand-painted hanafuda playing cards. The company would remain in the Yamauchi family for the next 78 years. By the mid-1960s the playing card business was sliding. The new president, Hiroshi Yamauchi, embarked on a frantic sequence of experiments (a taxi service, a chain of love hotels, a brand of instant rice), almost all of which failed. Then in 1977 the company made a move that nobody in the playing card industry had been expecting. What that move was, and what it eventually became, is the third thread. We will return to it.

For sixty years Fujifilm and Kodak were near-mirror competitors in chemical photography, each protected by patent thickets and capital intensity. Between 2000 and 2010 the entire global market for photographic film collapsed by ninety per cent as the smartphone arrived; Kodak filed for bankruptcy in 2012. Fujifilm did not, and the decision its CEO took in the middle of the collapse is the fourth thread.

By 2006 Netflix dominated the postal DVD-rental category, mailing more than a billion DVDs a year, and was the second-largest customer of the United States Postal Service. In 2007, when DVD-by-mail still produced ninety-nine per cent of revenue, Reed Hastings made a move that nobody in the rental industry was expecting. What that move was, and what it eventually cost the DVD business, is the fifth thread.

For thirty years Adobe sold Photoshop, Illustrator, and InDesign in shrink-wrapped boxes on retailer shelves, with a release cycle of roughly eighteen months and a margin profile any software CFO would defend with her life. In 2012, at the apex of that business, Adobe announced it would stop selling the boxed versions altogether and convert its entire flagship suite to a cloud-only subscription called Creative Cloud. The decision destroyed the firm’s most defensible revenue stream and lost it an enormous fraction of its installed customer base in eighteen months. What happened next is the sixth thread.

For thirty-nine years from its founding in 1975, Microsoft’s identity was a single word: Windows. In February 2014 a former cloud-engineering executive named Satya Nadella was appointed CEO, and his first internal memo did not contain that word. What he did next, against the firm’s own forty-year identity, is the seventh thread.

Marvel Entertainment filed for Chapter 11 bankruptcy in 1996 and, in the process of restructuring, had licensed away the screen rights to its most famous characters (Spider-Man, the X-Men, the Fantastic Four) to other studios in exchange for capital it desperately needed. By 2005 Marvel was left with the comparatively obscure properties no studio had bothered to acquire: Iron Man, Thor, Captain America, the Hulk. A move was then made that essentially nobody in Hollywood believed could work. What that move was, and what franchise it produced, is the eighth thread.

Founded in Bombay in 1942 by four chemists in a garage, Asian Paints by the 2010s held a market lead so wide that competitors had effectively conceded India. A paint business is, by every textbook metric, a defensible commodity, and the firm could have spent the decade defending the margin. It chose a different game, beginning with a question about what to do with the proprietary dataset of millions of colour-purchase decisions made by Indian households across three decades. What that question turned into is the ninth thread.

In 1968 the Tata Group set up a small computing unit to handle the back-office data processing of its own steel, motors, and chemicals divisions, entirely as a cost centre, with no commercial mandate, no external clients, and no industry positioning. Tata Consultancy Services would, across the next half-century, transform from this in-house support unit into the largest IT-services firm in India and one of the largest globally by revenue. The transformation involved at least three Anti-Company decisions, each made against the firm’s own settled identity and each made over the objections of its parent group. What those decisions were is the tenth thread.

Ten companies. Ten crises or strategic inflections. Ten signatures of a cognitive operation that the next section will name.

04The Grammar of Play — Eight Strategic Position Games

Play, in the sense this paper uses, is not a tone. It is not the team-building event with marshmallows and spaghetti. It is a structured cognitive operation that adults perform poorly because they have been rewarded for years for performing the opposite operation: judgment under scarcity, optimization within constraints, defence of what works. The imagination games below are scaffolds. They give a corporate group permission to perform the cognitive operation that de Mestral performed instinctively under his Swiss microscope, with a burdock burr and an Irish pointer and a fastener that was waiting in the geometry of a plant.

The taxonomy below distinguishes between the core three (the minimum kit any incumbent should be able to run within a quarter) and the extended five, which require more infrastructure but expand the repertoire of cognitive moves an organisation can perform on itself. None is a substitute for the others. Each games a different failure mode: assumption-blindness, peripheral-vision deficit, success-induced amnesia, signal-noise confusion, pattern-poverty, optimism over base rates, low-imagination forecasting, and the inability to climb the possibility tree one branch at a time.

—

SELECT A GAME

Tap any numbered node on the cycle to read its description, rule, outcome, real-world example, and failure mode.

1–2 · Raw imaginative material 3–4 · Structuring the field 5–6 · Operating on the firm 7–8 · Closing the loop

Figure 4 · The Strategic Position Games — An Eight-Step Cycle. The eight games arranged as a cycle, not a linear process. Reading clockwise: nodes 1–2 generate raw imaginative material (outlier signals and cross-industry templates); nodes 3–4 structure that material into competing futures and competing adversaries; nodes 5–6 operate on the firm itself (stress-testing the mirror image and enumerating the next move); nodes 7–8 close the loop by checking against the messages the periphery and the workforce are already quietly transmitting.

The Eight as a Sequence

The eight games are not a menu; they are a sequence with a logic. The Anomaly Game and the Analogy Game generate raw imaginative material: outlier signals and cross-industry templates. The Scenario Game and the War-Game structure that material into competing futures and competing adversaries. The Anti-Company Game and the Adjacent-Possible Game then operate on the firm itself, one stress-testing the mirror image, one enumerating the next move. The Maverick Game and the Pre-Mortem Game close the loop by checking the result against the messages the periphery and the workforce have already been quietly transmitting. An organisation that runs all eight in a calibrated annual rhythm is not a more imaginative version of an analytical firm. It is a structurally different organism. It is, to anticipate the next section, a firm that has built itself a playground.

These eight games together are an efficient near-complete grammar of play, but they are still not the whole grammar. They map cleanly onto the exploration-exploitation tension James March named: each one bends an organisation built to exploit back into a posture in which exploration becomes thinkable.^[13] What is striking about all eight is that not one of them requires capital expenditure, platform deployment, or headcount addition. The cost is psychological. Each game requires senior leaders to admit that the model that made them successful might be the model that destroys them. The fastener de Mestral pulled out of a burr is not whimsy. It is the executive function that the corporate body has surgically removed.

“The cost of play is not in dollars. It is in the willingness to admit that the model that made you successful may be the model that destroys you.”

05The Playgrounds — Environments Leadership Must Engineer

A grammar of play is necessary but not sufficient. A team that knows the eight games and tries to play any of them inside the standard operating cadence of a large incumbent firm (quarterly earnings discipline, decked-out steerco meetings, deliverable-driven sprints, performance-rated everything) will discover within a session that the games either dissipate into low-energy mimicry or get quietly relabelled as “ideation” and parked in the appendix of the slide deck. The games require a particular kind of operating environment to land. The literature on creativity, organisational learning, and complex-adaptive systems converges on a small set of design conditions for what I will call, throughout the rest of this paper, the playground. There are five of them.

1. Psychological Safety

Amy Edmondson's quarter-century of research at Harvard Business School is unambiguous: the single most powerful predictor of whether a team will surface uncomfortable signals, propose counter-orthodox ideas, and admit error early enough to course-correct is not talent, incentives, or process. It is the team's perceived psychological safety, the shared belief that taking interpersonal risks will not damage one's standing.^[14] Every imagination game above asks the player to say something that, in the prevailing professional paradigm, would be career-incongruent. The Anti-Company Game asks them to make the case for what their CEO has spent a decade refuting. The Pre-Mortem Game asks for what they suspect is true but cannot say. Without engineered psychological safety, every game collapses into the answer the room thinks the senior partner wants to hear.

2. Slack — Structurally Protected Time

The professional-services and modern corporate paradigm is to push utilisation as close to 100% as the body will tolerate. The playground requires the opposite: a structurally protected fraction of time (Google's 20%, 3M's 15%, the 3- to 5-day off-cycle “strategy retreat” in the regulated industries) during which the meter is not running. The protection must be structural, not aspirational. Time that managers have to fight for, justify, or recover by working evenings is not time the imagination operates in. The imagination operates only in the time it knows it owns.

3. Diversity Within the Room

Imagination, in the cognitive-science sense, is the recombination of distant elements. A room that is homogenous along dimensions of training, tenure, function, and worldview cannot generate the recombinations the games require; its element-set is too small. The playground requires deliberate composition: the strategist and the front-line operator, the engineer and the legal counsel, the lifer and the recent hire, the insider and the outsider. The discipline is not diversity for its own sake; it is diversity calibrated to the cognitive operation. The point of putting a frontline branch manager next to the head of strategy in the Pre-Mortem Game is precisely that the branch manager has the relevant data the head of strategy does not.

4. Tempo — Time-Boxed, Not Open-Ended

The playground is paradoxically disciplined. Open-ended creative time tends to dissipate; tight time-boxes concentrate it. Each of the eight games has a natural duration: the Anti-Company Game runs in two hours, the Anomaly Game in a half-day of preparation plus a 90-minute session, the War-Game in two days. The tempo is the constraint that makes play productive. The playground does not require leaders to find an extra week. It requires them to schedule the four hours, the half-day, the two days, and to protect those windows with the same operational discipline they apply to a board meeting.

5. The Galápagos Principle — Distance From the Core

Some playful work cannot survive proximity to the operating business. The mental model that runs the operating business is the model the playful work needs to challenge, and the gravitational pull of the operating model on any project run inside it is enormous. Successful imagination programmes at large firms show the pattern: Lockheed's Skunkworks (Kelly Johnson, 1943), IBM's Emerging Business Opportunities, Google X, and the “ambidextrous organisation” design championed by O'Reilly and Tushman^[15] . All of them locate playful work in a structurally separate unit with its own funding, its own metrics, and its own physical and cultural distance from the core. The principle is borrowed from evolutionary biology: novel species evolve most readily in geographically isolated populations, such as the Galápagos finches and the Madagascar lemurs, where the dominant mainland gene pool cannot keep recolonising and reabsorbing the variant.

Amy Edmondson’s quarter-century of research at Harvard Business School is unambiguous: the single most powerful predictor of whether a team will surface uncomfortable signals, propose counter-orthodox ideas, and admit error early enough to course-correct is not talent, incentives, or process. It is the team’s perceived psychological safety, the shared belief that taking interpersonal risks will not damage one’s standing. Every imagination game above asks the player to say something that, in the prevailing professional paradigm, would be career-incongruent. The Anti-Company Game asks them to make the case for what their CEO has spent a decade refuting. The Pre-Mortem Game asks for what they suspect is true but cannot say. Without engineered psychological safety, every game collapses into the answer the room thinks the senior partner wants to hear.

Real-World Examples

Google’s Project Aristotle (2012–2015)

Google’s People Analytics team studied 180+ internal teams over two years to identify what made teams effective. The variable that most strongly predicted performance was not member talent, seniority, workload, or shared goals; it was psychological safety. Teams with high psychological safety took on riskier projects and surfaced uncomfortable observations more often. They also recovered faster from setbacks. The finding reshaped Google’s manager-training programmes and was published in 2016 to broad public reception.

Pixar’s Brain Trust (2003–present)

Ed Catmull’s deliberately designed review ritual gathers senior directors (Lasseter, Stanton, Docter, Bird, and rotating peers) to give unsparing feedback on each film-in-progress. The rule that protects the honesty: nobody in the Brain Trust has authority to mandate changes; only the film’s director can. The architecture removes the political risk of being told the work isn’t good enough while preserving the director’s creative ownership. Every Pixar film from Toy Story through Up passed through this room.

The professional-services and modern corporate paradigm is to push utilisation as close to 100% as the body will tolerate. The playground requires the opposite: a structurally protected fraction of time (Google’s 20%, 3M’s 15%, the three- to five-day off-cycle “strategy retreat” in the regulated industries) during which the meter is not running. The protection must be structural, not aspirational. Time that managers have to fight for, justify, or recover by working evenings is not time the imagination operates in. The imagination operates only in the time it knows it owns.

Real-World Examples

3M’s 15% Time (1948–present)

3M’s institutional policy that researchers spend 15% of working time on self-directed projects has produced the Post-it Note (Spencer Silver’s “low-tack” adhesive sat unused from 1968 until Art Fry recombined it with a bookmarking idea in 1974; the product launched in 1980) and a long tail of patents in adjacent and unrelated categories. The 15% is structurally protected from quarterly performance review; researchers do not have to justify it.

Google’s 20% Time (1998–c. 2013)

The founding-era engineering policy that allowed 20% of work hours on self-chosen projects produced Gmail (Paul Buchheit’s 20% project, launched 2004), AdSense, and Google News (Krishna Bharat). The discipline began eroding around 2013 as quarterly metrics tightened; the products born of the 20% era still drive significant portions of Google’s revenue today.

Real-World Examples

Apollo 13’s CO₂ Scrubber Rescue (April 1970)

When the Apollo 13 lunar module’s CO₂ scrubbers failed at hour 60 of the mission, NASA assembled a 12-person team across mechanical engineering, life-support systems, materials science, EVA suit design, and mission operations. They constructed a working scrubber from available spacecraft parts (a sock, duct tape, the cover of a flight manual, a hose) and radioed the procedure up to the crew. The solution worked within three hours of being communicated. A homogeneous team would not have known what was available aboard the spacecraft to repurpose.

Eli Lilly’s InnoCentive (2001–present)

Eli Lilly’s R&D innovation platform posts unsolved chemistry and biology problems to a global network of solvers from any field. A study by Karim Lakhani (Harvard Business School) of 166 problems posted between 2001 and 2006 found that the probability of a successful solution rose with the cognitive distance between the solver’s discipline and the problem domain. Chemistry problems were most often solved by people from non-chemistry fields (physicists, biologists, materials engineers) who recombined the problem against frames the chemists had stopped seeing.

Real-World Examples

Amazon’s Six-Page Narrative Memo (2004–present)

Jeff Bezos banned PowerPoint at Amazon executive meetings in 2004 and replaced it with the discipline of a six-page narrative memo, read in silence for the first 20 minutes of every meeting before discussion begins. The format forces the writer to think through arguments rather than rely on bullet-point compression. The tempo, 20 minutes of silent reading and then discussion, is rigid and identical across every Amazon leadership meeting.

Google Ventures Design Sprint (2010s)

Jake Knapp at Google Ventures developed the five-day time-boxed innovation method (Map → Sketch → Decide → Prototype → Test) used to compress what would normally be months of product strategy into a single working week. The discipline is the constraint: a sprint that exceeds five days isn’t a sprint. Documented use at Slack, Blue Bottle Coffee, Savioke, and across the GV portfolio.

Some playful work cannot survive proximity to the operating business. The mental model that runs the operating business is the model the playful work needs to challenge, and the gravitational pull of the operating model on any project run inside it is enormous. Successful imagination programmes at large firms show the pattern: Lockheed’s Skunkworks, IBM’s Emerging Business Opportunities, Google X, and the “ambidextrous organisation” design championed by O’Reilly and Tushman. All of them locate playful work in a structurally separate unit with its own funding, its own metrics, and its own physical and cultural distance from the core. The principle is borrowed from evolutionary biology: novel species evolve most readily in geographically isolated populations, such as the Galápagos finches and the Madagascar lemurs, where the dominant mainland gene pool cannot keep recolonising and reabsorbing the variant.

Real-World Examples

Lockheed Skunkworks (1943–present)

Kelly Johnson established Lockheed’s Advanced Development Programs unit as a structurally separate operation, physically located off the main Lockheed campus, with its own funding, hiring rules, and a 14-rule operating manual (e.g., “the project manager must be delegated practically complete control”). The unit shipped the P-80 Shooting Star in 143 days (1944), the U-2 reconnaissance plane (1955), the SR-71 Blackbird (1964), and the F-117 stealth fighter (1981), each impossible to produce inside Lockheed’s main engineering organisation.

Apple’s iPhone “Project Purple” (2004–2007)

Scott Forstall recruited the iPhone software team onto a locked floor at Apple’s Infinite Loop campus, with badge-controlled access, where members were forbidden from telling family or colleagues what they were working on. The team operated as a startup inside Apple, with its own design language and engineering velocity, isolated from the iPod division it would eventually displace. The iPhone shipped in January 2007 against a culture (iPod) that the project team had to be structurally separated from to challenge.

Figure 6. The five design conditions of a playground. Each constraint is necessary; none is sufficient. Tap any node above to read the full description and real-world examples; tap back to switch.

What leadership owes the playground, in the end, is not enthusiasm. Enthusiasm is the easiest of the five conditions and the least durable. What leadership owes the playground is the structural protection of the other four: psychological safety preserved when the result is unwelcome, slack defended when utilisation is under quarterly pressure, diversity protected when promotion processes drift toward homogeneity, and Galápagos distance maintained when the parent business demands that the playful unit conform to its operating cadence. The leader's role in the playground is custodian, not protagonist.

06Reading the Outcomes — From Play to Deployment

Suppose the leadership has built the playground and run the games. What comes out? The output of play is not a strategy document. It is a portfolio of candidate signals : possibilities, anomalies, mirror-image futures, mavericks worth tracking, each of which sits somewhere on a spectrum between “interesting curiosity” and “material commitment.” The most common failure mode of corporate imagination programmes is not the inability to generate the signals. It is the inability to read them. The playground produces material that the operating organisation has not been trained to evaluate. Three reading disciplines bridge the gap.

Reading Discipline 1 — Distinguishing Signal from Noise

The Anomaly Game in particular generates a high volume of outliers, most of which are noise. The discipline is to apply structural tests, not narrative tests. A narrative test asks whether the anomaly is an interesting story; almost every anomaly is. A structural test asks: does the anomaly recur, does it correlate with an exogenous driver, does it concentrate within a definable population, does the unit economics around it move in the right direction? The Reeves and Goodson protocol described in their 2021 Harvard Business Review piece on anomaly-driven strategy^[5] formalises this into a four-question filter: (i) is the anomaly persistent? (ii) is it spreading? (iii) does it have a recognisable causal mechanism? (iv) can we be early relative to the broader market? An anomaly that passes all four warrants investment; one that passes three warrants a watching brief; one that passes fewer goes back in the bin without guilt.

Reading Discipline 2 — Real-Options Capital Allocation

A signal that has passed the filter does not yet justify a full strategic commitment. The right deployment vehicle is the real option: a small, time-bound investment that buys the right (but not the obligation) to scale if the underlying uncertainty resolves favourably.^[16] Each game's output maps to a particular real-option shape. The Anti-Company Game typically yields a defensive option (a small business unit running the mirror-image model, as a hedge). The Maverick Game typically yields a partnership or investment option (a minority stake, a JV, a commercial alliance with the maverick firm). The Pre-Mortem Game yields what is best thought of as an insurance option (an internal change programme purchased against an identified failure mode). The discipline is to size the option correctly relative to the underlying volatility: large enough to learn from, small enough that the firm survives being wrong, and structured so that the next funding decision is gated on a clearly specified piece of new information.

Reading Discipline 3 — The Portfolio Logic

A single playful output does not survive a hostile budgeting environment. A portfolio of them might. The leadership discipline is to maintain a small but durable set of live options, perhaps ten to twenty across the firm, ranging from early-stage curiosity (a watching brief on a maverick) to advanced-stage commitment (a separate business unit), and to manage the portfolio with explicit failure budgets, kill-decision criteria, and graduation milestones. The portfolio approach inoculates the imagination programme against two failure modes. First, it prevents the imagination unit from being judged by the success of any single bet (most bets will fail; the portfolio's expected value depends on the rare large winner). Second, it prevents the strategy of the firm from being captured by any single playful insight (which would convert the playground into another orthodoxy).

From Play to Deployment — The Convergence Move

The hardest move in the cycle is not divergence (the playground does this naturally) but convergence: the moment a signal that has cleared the filter, the real-options sizing, and the portfolio gate must become an operational commitment of the main organisation. This is where most playful programmes break. The operating organisation either rejects the import as foreign, or absorbs and dilutes it into the existing operating model. The literature on ambidextrous organisations^[15] offers a clean design solution: senior leadership owns both the exploration unit and the exploitation unit, sits structurally between them, and is explicitly compensated on the long-run performance of the portfolio rather than on the short-run efficiency of the operating business. Without this organisational design, the convergence move fails by default, and the playground becomes a budget line item with no causal connection to the firm's strategic trajectory.

“The playground generates possibilities. The reading disciplines turn possibilities into options. The convergence move turns options into commitments. Most firms can do the first; few can do the second; almost none can do the third.”

07The Agentic AI Reframe — Why the Playground Becomes Existential

The entire argument above can be read as old wine. Strategists have argued for imagination, divergent thinking, and creativity since at least the 1960s; the games have been documented since the 1980s; the playground design conditions have been studied since the 1990s. What has changed, and what makes the case in this paper materially different from any prior version of it, is the arrival of frontier AI systems that match or exceed human performance on the two cognitive modes that classical management has rewarded most: deduction and induction.

The Three Modes of Reasoning

The American philosopher Charles Sanders Peirce, writing in the late nineteenth century, distinguished three modes of logical inference.^[17]Deduction derives necessary conclusions from given premises (all swans are birds; this is a swan; therefore this is a bird). Induction generalises from observed instances to provisional rules (I have seen ten thousand white swans; swans are probably white). Abduction (Peirce's term, sometimes called retroduction or inference to the best explanation) generates a hypothesis that, if true, would explain a surprising observation (a black swan! Either my generalisation was wrong, or this is not a swan, or something is different about Australia). Of the three, Peirce was emphatic that abduction is the only one that introduces genuinely new ideas. Deduction unpacks what was already implicit. Induction extracts the predictable from the seen. Abduction proposes what might be true but is not yet observable in the data.

Where Frontier AI Wins, and Where It Does Not

Frontier large language models, by 2026, perform deductive and inductive operations at the level of a competent senior analyst, and increasingly above it. The deductive operation that took a junior associate three days (extending a financial model under given assumptions; deriving the regulatory implication of a new clause; computing the cascade through a complex consolidation) is now performed in minutes. The inductive operation that filled a quarter (pattern-extraction across a corpus of transactions, contracts, support tickets, or audit findings) is performed in hours. The compression is empirical and ongoing.

The abductive operation is different. A growing body of research has shown that frontier LLMs underperform humans systematically on tasks requiring true abductive reasoning: Ribeiro and Carmona (2025) in ScienceDirect^[18], the systematic survey by Liu et al. (2026) in Transactions on Machine Learning Research^[19], and the syllogistic-form work by Yamamoto et al. (2026)^[20] . The reason is structural, not a matter of model scale. LLMs are trained on text that records what humans have already thought; they extrapolate the distribution of what has been said. The abductive operation requires reasoning about what is true but has not yet been said, about hypotheses that would explain surprising data, where the relevant surprising data is itself out-of-distribution. Lipton's distinction between lovely explanations (those that satisfy human preferences for coherence and elegance) and likely explanations (those that correctly identify actual causes) is operative here.^[21] LLMs are optimised, by their training objective, to produce the lovely. The likely is harder, and on novel terrain remains stubbornly human.

Figure 7. The three modes of reasoning and where frontier AI stands as of 2026. Deduction and induction, the two operations classical management has rewarded most, are now matched or exceeded by frontier LLMs. Abduction, the cognitive operation that generates genuinely new hypotheses from surprising data, remains a structural human advantage. The eight imagination games above are, almost without exception, abductive exercises in disguise.

The Strategic Consequence

If you trace the cognitive operations that the eight imagination games actually perform, the pattern is striking. The Anti-Company Game generates an abductive hypothesis about a mirror-image firm. The Anomaly Game generates abductive hypotheses about what an outlier might be an example of. The Analogy Game generates abductive hypotheses about which distant industry's mechanism would explain the present situation. The Maverick Game generates abductive hypotheses about what the maverick might be seeing that the incumbent is not. The Pre-Mortem Game generates abductive hypotheses about which mechanism would kill the firm. The Scenario Game, the War-Game, and the Adjacent-Possible Game each generate abductive hypotheses about possible futures, possible adversary moves, and possible next steps. The grammar of play is, almost without exception, a grammar of abduction.

This collapses the entire imagination programme into a sharper strategic claim. A company whose strategic moat depended, until 2024, on the relative quality of its deductive and inductive analytic capability is now a company whose moat is being arbitraged away in real time by frontier AI systems available to its competitors at a marginal cost approaching zero. The deductive workflow the consulting firm sold for $400,000 in 2019 is now a $4,000 LLM-assisted exercise. The inductive pattern recognition that the audit firm sold for $1.2 million is now a multi-day automated run. The human contribution that remains scarce, durable, and strategically priced is the abductive operation. And the only known organisational technology for generating abductive operations reliably and at scale is the playground.

Three Implications for the Firm

(i) The playground is no longer optional. A firm that does not run abductive operations is a firm whose only remaining differentiator is the speed at which it can apply commoditised deductive and inductive AI capability, which is to say no differentiator at all in any time horizon long enough to matter. The argument shifts from “imagination is a useful supplement to analysis” to “imagination is the analysis that remains scarce, and therefore priced.”

(ii) Human capital is repriced. The professionals who remain valuable in the agentic-AI era are those who can run abductive operations. Classical management has not been selecting for that profile. The deductive specialist becomes a partial substitute for the LLM; the abductive generalist becomes scarce. Talent strategy, performance management, partner selection, and educational pipelines all need to shift accordingly. The most expensive misallocation a firm can make in the next decade is to continue recruiting and rewarding the deductive specialist as if her cognitive operation were still rare.

(iii) Agentic AI is the new instrument in the playground, not its replacement. The interesting design question is not whether to use frontier AI in the playground (refusing to is professional malpractice) but how. The right use is asymmetric. Use AI to do the inductive sweep across the entire customer base in search of anomalies; use the human team to do the abductive operation on the outliers the AI surfaces. Use AI to run thousands of scenario branches; use the human team to identify which branches encode genuinely novel mechanisms versus which are recombinations of the training data. Use AI as the indefatigable War-Game adversary that can play all night; use the human team to interpret what the surviving strategies have in common. The playground is augmented by AI, not replaced by it. The asymmetric pairing is the new operating manual.

“A company without a playground in the agentic-AI era is selling deductive and inductive labour against a frontier model whose marginal cost is approaching zero. The half-life of that business is the half-life of the next model release.”

08Playfulness for the Consulting Firm

Now turn the lens on the firms reading this paper. The major consulting firms are not exempt from the half-life argument: the strategy consultancies, the management-consulting arms of the Big-4 (Deloitte Consulting, EY-Parthenon, KPMG Strategy, PwC Strategy&), and the implementation giants (Accenture, IBM Consulting, Capgemini). They are its central case. The product they sell is the application of intelligence to ambiguous problems, and that product is now being delivered, increasingly, by AI systems running over the same data the consultant used to interpret. The strategy diagnostic that filled a quarterly engagement is being assembled by clients themselves from open-weights models and a half-day of LLM-assisted analysis. The multi-year EPM implementation is being compressed by pre-configured industry templates and AI-assisted blueprinting. The half-life of consulting advantage is collapsing on the firms that built their careers naming the same collapse in their clients.

The orthodox response (deeper specialization, faster delivery, AI augmentation, asset productisation) is the exploitation response. It is necessary, it will continue, and on its own it will lose. The exploitation move has to be paired with an exploration move, which means consulting firms need to relocate themselves on the edge-of-chaos quadrant in Figure 2. What follows is what that looks like, practice by practice.

Figure 5. Playfulness across consulting practice domains: incumbent norm vs. playful potential. The shape of the incumbent figure is constrained by audit independence rules, regulatory exposure, and the deliverable-discipline culture of large firms. The shape of the playful potential is what becomes possible when the three imagination games are systematized into the engagement model itself, not bolted on as offsite theatre.

Strategy — The Anti-Company Game in the Room

The strategy practices already host more playful infrastructure than they admit. Deloitte's Greenhouse, McKinsey's Studio, EY's wavespace, the in-house innovation labs at the strategic consultancies: each is a physical space designed to break the cadence of the client's normal decision-making. The infrastructure exists. What is often missing is the engineered playful exercise within the space. The Anti-Company Game is the most direct fit: in a half-day session, the client team writes down its strategic certainties, the three to five sentences they would put on a board slide describing what the firm is for, and then constructs the best case for the opposite firm. The consultant's job in this room is not to facilitate consensus. It is to refuse to let the team converge on the mirror image too quickly. The most valuable client minute is the seventy-third one, when the room realises the mirror-image firm is not absurd. It is the firm a competitor is already building.

EPM Implementation on SAP and Oracle — The Playful Blueprint

Implementing an enterprise performance management platform (SAP S/4HANA with SAP Group Reporting, or Oracle Fusion with EPM Cloud) is the practice domain that looks furthest from play. It is engineering. It is configuration. It is documentation. The deliverable is a working system, and working systems are not built by holding incompatible elements in productive tension. They are built by reducing tension. The deepest mistake in this practice is to believe that statement. The most expensive EPM implementations are the ones that translate the client's current operating model into the new platform without ever asking whether the current operating model is the right one. The chart of accounts, the cost-centre hierarchy, the planning calendar, the consolidation logic: these are all artefacts of the firm the client was, not necessarily the firm the client wants to become.

The playful blueprint workshop does to an EPM implementation what the Anti-Company Game does to strategy. In the blueprint phase, the client team is asked to construct three versions of the future state: the literal translation of the current model into the new platform, a mirror-image model in which every sacred design choice is inverted, and a maverick model imagined as if a digital-native upstart in the client's industry had built the EPM stack from scratch. The implementation team is then asked which design decisions show up in only one of the three blueprints (the points of fragility) and which show up in all three (the points of genuine invariance). The platform configuration that follows is no slower than the orthodox approach. It is, repeatedly, two to three years more durable, because it has been stress-tested against alternative futures rather than calibrated to the present. For a five-year implementation cycle in an industry whose half-life of advantage is one year, the playful blueprint is not a soft addition. It is the only honest design discipline left.

09Resolution — The Threads Pull Together

It is time to close the threads.

The LEGO Group in 2003 did not solve its crisis by diversifying further. Knudstorp did the opposite. He sold the theme parks. He cut the video game arm. He ended the apparel licences. He returned the company to a single proposition: the brick, and the human being playing with it. The decision looked, at the time, like a strategic retreat. It was in fact the most aggressive Anti-Company move imaginable: a flip of every sacred 1990s assumption about what a modern toy brand had to be. Knudstorp opened the design process to Adult Fans Of LEGO, crowdsourced new product lines through what became the LEGO Ideas platform, and let the Mindstorms robotics community continue modifying the firmware of his own products. By 2024 the LEGO Group was the most valuable toy brand in the world and outsold every competitor in its category, in many years by a wide margin. Its turnaround was not a return to first principles. It was a return to play.

The Pixar Brain Trust is a no-rank room. A film in development is brought to the room when its director feels it needs help. The Brain Trust (Catmull, Lasseter, Stanton, Docter, Bird, and a rotating cast of peers) watches an unfinished cut and then, for hours, says everything that is wrong with it. There is one rule: nobody in the room has authority to order a change. The director listens, fights, refuses, accepts, departs. Every Pixar film from Toy Story through Up , and most of them since, passed through that room. The Brain Trust is the institutional embodiment of the Pre-Mortem Game: the room exists to make the obituary thinkable before the film is finished, while there is still time to rewrite it. The Brain Trust did not always succeed. But the films that struggled most after release were almost always the ones the Brain Trust had been politely talked out of dismantling.

The Nintendo move of 1977 was the Color TV-Game. It was a home video game console, a domain in which Nintendo had no track record, no engineering competence, and no industry relationships. It sold three million units. Hiroshi Yamauchi had, in effect, declared the 88-year-old hanafuda business obsolete and made the leap. Eight years later the company released the Nintendo Entertainment System, with a moustachioed plumber named Mario as its mascot. Twenty-one years after that, in 2006, it released the Wii, a console that competed not on raw graphics but on the playful act of physical motion. In 2017 it released the Switch, which collapsed the console-handheld distinction and outsold every prior Nintendo platform. Each pivot was a Maverick move executed by an incumbent against itself: Nintendo refusing to be the miniature version of the company it had been the year before.

Fujifilm's move was made by a CEO named Shigetaka Komori between 2000 and 2010, in the same chemical-photography market that killed Kodak. Komori took a perverse decision early: he refused to treat the company's chemical-film expertise as an obsolete asset, and asked instead what else, structurally, the same expertise could do. The gelatin coating on photographic film, his scientists pointed out, shared chemistry with the proteins that age the human skin. The anti-oxidation problem in long-archival film was the same problem as cellular oxidation in cosmetics. From this Anti-Company reasoning emerged Astalift, Fujifilm's now-billion-dollar skincare line; a pivot into medical imaging through the Synapse and REiLI platforms; an acquisition of Toyama Chemical that placed Fujifilm in pharmaceuticals; and a regenerative-medicine subsidiary working on iPS cells. By 2024 the chemical-photography business, once 60 per cent of revenue, was under five. The company had not died with its market; it had translated.

Netflix's move was even more aggressive. In 2007, when DVD-by-mail still produced ninety-nine per cent of revenue and the streaming bandwidth most American homes possessed could barely sustain a single low-resolution video, Reed Hastings authorised the launch of the streaming service that would, within a decade, render the DVD business obsolete. The Anti-Company move was made against the firm's own most profitable line, not against an external competitor. The Pre-Mortem reasoning was explicit: Hastings later wrote that the company had performed the funeral on its DVD business in 2007 to give itself eight years of head-start in the business that would replace it. By 2013 Netflix produced House of Cards, its first original series, executing the Maverick move yet again, this time against the entire studio system that had previously supplied its catalogue. The half-life of each Netflix advantage has been short; the company's competence at performing the Anti-Company Game on itself has not been.

Microsoft's resolution is the most recent. When Satya Nadella took over as CEO in February 2014, the company's identity was “Windows everywhere”: Windows on the PC, Windows Phone in your pocket, Windows in the server room. Nadella's first internal memo announced that the strategy was now “mobile-first, cloud-first”. The omission of the word “Windows” was deliberate. Within three years Microsoft had open-sourced .NET, made Office natively available on iOS and Android, embraced Linux on Azure (a sentence that would have been a sackable offence in the Ballmer era), and reorganised the company around cloud rather than the operating system. The Anti-Company Game had been played on the firm's own forty-year identity, and the firm had survived its own erasure. By 2024 Azure was the second-largest cloud platform in the world and Microsoft's market capitalisation had risen by a factor of nine.

Asian Paints, founded in Bombay in 1942, is the most counterintuitive case in the set. A paint company has, by every textbook measure, a defensible commodity business: production scale, distribution reach, brand. It has no obvious adjacent possible. And yet across a decade Asian Paints quietly used its proprietary trove of millions of colour-purchase decisions, the most granular national dataset of Indian aesthetic preference in existence, to launch ColourNext, a trend-forecasting platform; Beautiful Homes, a home-décor and interior-design service; and an ecosystem of installation, consultation and visualisation services that converted a paint company into the de-facto designer of urban Indian residential interiors. The strategic question Asian Paints kept asking was the Adjacent-Possible question: given what we already see about how people live, what is one move away from selling paint? The cumulative answer turned a commodity business into a platform, without ever leaving the paint shelf at the front.

Adobe's move was the cleanest Anti-Company decision in modern software. At the apex of its boxed-software business in 2012, the company announced that all future versions of Photoshop, Illustrator, and InDesign would be available only through a cloud subscription called Creative Cloud; the shrink-wrapped box would be discontinued. The market reaction was uniformly hostile: a Change.org petition gathered fifty thousand signatures from professional designers, the financial press wrote obituaries, and Adobe lost a meaningful slice of its installed base within eighteen months. The Pre-Mortem reasoning had been explicit inside the firm: a fixed-release shrink-wrapped product, the executives believed, would be impossible to defend by 2018 against piracy, free-tool competitors, and the structural advantages of subscription incumbents. By 2020 Adobe's subscription revenue had crossed the boxed-revenue peak by a factor of three; by 2024 it had crossed it by a factor of eight. The boxed Photoshop is a museum exhibit; Creative Cloud is the industry standard.

Marvel's move was made between 2005 and 2008, by a chief operating officer named David Maisel and a chief creative officer named Kevin Feige, neither of them a Hollywood power at the time. With the firm's most famous characters licensed away to other studios, Marvel chose to mortgage virtually every remaining property as collateral and finance, in-house, a film centred on a second-tier hero (Iron Man), to be directed by a comedy filmmaker (Jon Favreau), starring an actor whose career was widely considered finished (Robert Downey Jr.). The Adjacent-Possible move was not the film itself. It was the explicit decision to build, from the start, an inter-connected story universe in which each film fed customers into the next, releasing characters from individual-franchise mortality. The Marvel Cinematic Universe, by 2024, had passed thirty films and over thirty billion dollars in cumulative box office, becoming the highest-grossing franchise in cinema history. The firm walked out of bankruptcy with the wrong characters and produced the most valuable film franchise ever made.

The Tata Group's move was three Anti-Company decisions stacked on top of each other across half a century. The first was in 1968: F.C. Kohli, the head of the in-house Tata Consultancy Services unit, persuaded the Tata Sons board to allow him to sell the unit's services to clients outside the Tata Group, an unprecedented violation of the captive-back-office model. The second was in 1974 when, against vocal internal opposition, Kohli signed a partnership with Burroughs Corporation to handle programming work shipped from the United States, at a time when Indian regulation made even foreign-exchange reporting an ordeal. The third was through the 1980s and 1990s when TCS rebuilt its operating model around the “Global Network Delivery Model”, executing engagements simultaneously across geographies in a way that no Western competitor had constructed. By 2024 Tata Consultancy Services had over six hundred thousand employees, a market capitalisation north of one hundred and fifty billion dollars, and a global services franchise the captive-IT model of 1968 was not capable of imagining. Each anti-company move was opposed at the time it was made by people the firm had every reason to believe.

Ten threads. Ten resolutions. Each company survived its existential moment by performing the cognitive operation de Mestral performed under his microscope. They held the sacred and the absurd in the same room long enough to see what they could make together.

Which returns us, finally, to the burr. It is not an inventor's parlour trick. It is a syntactical signature: the moment a mind decides that a thing built from a Swiss meadow and a microscope and a botanical seed-head is allowed to exist as a fastener long enough to find out whether it works. That signature is the deep grammar of re-imagination. It is what the LEGO Group recovered when it sold the theme parks. It is what Pixar built a room to host. It is what Nintendo performed five times in eighty-eight years. It is what Fujifilm assembled from the chemistry of film. It is what Netflix performed on its own most profitable line. It is what Adobe did to the boxed-software business it had spent twenty years perfecting. It is what Microsoft did to the founding identity of its founder. It is what Marvel did after walking out of bankruptcy with no characters left to license. It is what Asian Paints saw in the data on the colour of an Indian living-room wall. It is what the Tata Group built when it turned its in-house IT department into a global services champion. It is what every consulting firm reading this paper will need to manufacture deliberately, because the conditions that previously made it accidental (slack, tenure, optionality, time) have been engineered out of professional services delivery in the name of efficiency. The cost of play is now an explicit budget line, because the half-life of every advantage you have just sold a client is one year, and the half-life of the advantage that lets you sell it is on the same clock.

Author's Lens

Key Insights

Six portable observations from the synthesis. Each card holds an insight a reader could carry into a different setting (a board meeting, a client room, a portfolio review) and still recognise the shape of.

Half-Lives Have Made Re-Imagination the Primitive Move

Sustained competitive advantage no longer compounds across decades; it amortises across months. The implication is structural, not motivational. Every operating model assumes a temporal contract with the future, and a contract built on a ten-year half-life is mispriced by an order of magnitude when the half-life is one year.

Play Is the Cognitive Precondition, Not the Cultural Flourish

Play, in the sense this paper uses, is not the tone of an offsite or the mood of a workshop. It is the deliberate cognitive operation of holding incompatible elements in productive tension long enough to discover what they make. It precedes strategy, which is why strategy without it produces variations on the current model rather than alternatives to it.

The Edge of Chaos Is a Strategic Position, Not a Personality

Firms do not occupy the edge of chaos by hiring playful people. They occupy it by engineering rituals (the anti-company game in the strategy room, the pre-mortem in the risk review, the maverick scan in the engagement letter) that make the playful operation a scheduled deliverable. Personality is volatile; ritual is durable.

The Most Dangerous Spreadsheet Is the One That Was Right Last Quarter

Past success is the most reliable manufacturer of mental models that fail silently when the environment changes. The signal that a firm is approaching the cliff is not deteriorating performance; that signal arrives too late to be useful. The signal is the increasing internal cost of disagreeing with a current operating assumption. When dissent gets expensive, the cliff is near.

Consulting Is the Practice Most Exposed to Its Own Diagnosis

The firm that sells re-imagination to its clients while exploiting its existing delivery model is structurally identical to the brick-and-mortar bank that watched its branches empty without re-imagining its product. The consulting industry is being eroded on the same vector it advises clients to defend on, and the asymmetry of vision, clear about clients and opaque about self, is the most expensive blind spot in professional services today.

Play Is Procurable, but Only If It Becomes a Line Item

Slack, tenure, optionality, time: the conditions that previously made playful re-imagination accidental in mature firms have been engineered out of professional services delivery in the name of efficiency. They can be re-engineered in, but only as deliberate budget lines, not as a hope that someone will think laterally between Teams calls. The line item is itself the strategic move.

Actionable Takeaways

Six concrete moves the reader can run from this paper into next week's work. The EXPAND column is for the curious reader who wants to go deeper; the CONTRACT column is for the operator who needs a portable mnemonic.

EXPAND Go deeper

Read: Reeves & Fuller, The Imagination Machine (HBR Press, 2021) for the original game design and case studies.
Try: Block a half-day with your top team. List your five sacred operating assumptions. Build the best case for the firm that holds the mirror image.
Ask: If a competitor were running this exercise on us today, which of our assumptions would they invert first?
Connect: March's exploration-exploitation framework (1991) is the academic parent of this exercise; re-read it in light of your own org chart.

CONTRACT Remember & apply

Principle: The mirror-image firm is not absurd; it is the firm a competitor is already building.
Metaphor: Look at your strategy through a mirror, then ask which image is the real one.
Rule: Once a year, the executive team writes the case for the firm it has just chosen not to be.
Trigger: Whenever the senior team agrees with itself unusually quickly, schedule the Anti-Company Game within fourteen days.

EXPAND Go deeper

Read: Klein, G. (2007) “Performing a Project Premortem,” Harvard Business Review — the methodological antecedent.
Try: At the next annual risk review, ask the team to assume the firm has experienced a catastrophic failure five years out. Draft the post-event regulator's report.
Ask: What do employees already suspect but feel cannot be said while the current model is profitable?
Connect: Catmull's Pixar Brain Trust is the institutional embodiment of a continuous pre-mortem. Read it alongside Klein for the cultural side.

CONTRACT Remember & apply

Principle: The obituary written early is the obituary that does not have to be filed.
Metaphor: A pre-mortem is an early-warning system disguised as a death certificate.
Rule: Every annual risk review concludes with one written pre-mortem scenario authored by the team itself.
Trigger: Whenever a quarterly review begins with the phrase “our business model is fundamentally sound,” schedule the next pre-mortem within thirty days.

EXPAND Go deeper

Read: Christensen, C. (1997) The Innovator's Dilemma — the canonical treatment of why incumbents under-respond to disruptive entrants.
Try: For each major client, list five small, odd competitors whose business model the client's own management cannot fully explain. Refresh quarterly.
Ask: Suppose each of these mavericks has the right bet on the future, what does that imply for the client's three-year strategic plan?
Connect: Brooks Automation's pivot from semiconductor handling to biobanking shows the maverick scan applied to one's own business, not just a client's.

CONTRACT Remember & apply

Principle: The most accurate forecast of an incumbent's future is the business model of its smallest, oddest competitor.
Metaphor: On the edge of every industry is a population of strange firms. They are not noise; they are the next signal.
Rule: Every engagement file contains a refreshed list of five mavericks the client cannot quite explain.
Trigger: Whenever a client says “those firms are not real competitors,” add them to the watch list within the same week.

EXPAND Go deeper

Read: Reeves, M., Levin, S., & Ueda, D. (2016) “The Biology of Corporate Survival,” HBR — for the systems-biology logic of resilient design.
Try: In the blueprint phase of the next SAP S/4HANA or Oracle EPM implementation, construct three versions: literal, mirror-image, and maverick. Compare design decisions across all three.
Ask: Which configuration choices appear in only one blueprint, and which appear in all three?
Connect: Kauffman's edge of chaos applies as cleanly to a chart-of-accounts design as to a biological system; both are searches for the most adaptive minimum structure.

CONTRACT Remember & apply

Principle: An EPM blueprint that translates the current model into the new platform is a five-year bet that the current model is correct.
Metaphor: Three blueprints, side by side, are a wind tunnel for your finance architecture.
Rule: No EPM platform goes to design-freeze without three blueprints on the wall.
Trigger: Whenever someone says “let us just lift-and-shift the current chart of accounts,” open a second blueprint within forty-eight hours.

EXPAND Go deeper

Read: March, J. (1991) “Exploration and Exploitation in Organizational Learning,” Organization Science 2(1) — for why explore-budgets get silently starved.
Try: Allocate a named percentage of each practice's billable capacity, even five percent, to playful infrastructure: anti-company sessions, pre-mortems, maverick scans. Track it as you track utilization.
Ask: What is the dollar value of the playful capacity we are silently using up by leaving it unbudgeted?
Connect: O'Reilly and Tushman's ambidextrous organization framework (2004) is the structural design for this: separate playful units, common executive sponsorship.

CONTRACT Remember & apply

Principle: What is not in the budget is what gets eaten by the budget.
Metaphor: Play without a line item is a garden without a fence.
Rule: Every consulting practice plan contains an explicit play-capacity line, in basis points of total capacity.
Trigger: At every quarterly P&L review, the leadership team reads the play-capacity number aloud and explains its movement.

EXPAND Go deeper

Read: Kauffman, S. (1995) At Home in the Universe — the original treatment of the edge-of-chaos regime.
Try: Once a year, position your firm on the 2×2 in Figure 2. Use external data (partner attrition by tenure, time-to-promotion variance, client retention by tenure of relationship) to triangulate the dot's location.
Ask: If our firm's dot has not moved upward or rightward in three years, what is keeping it in place?
Connect: Complex adaptive systems theory predicts that systems too far from the edge of chaos either ossify or dissipate; neither is a survivable state for a knowledge business.

CONTRACT Remember & apply

Principle: A firm that does not know where it sits on the edge-of-chaos quadrant is not making a strategic choice; it is being moved by external currents.
Metaphor: The edge of chaos is not a place you find; it is a place you re-locate to every year.
Rule: Every annual partner retreat opens with a re-positioning of the firm's dot on the edge-of-chaos 2×2.
Trigger: Whenever the executive committee debates a major investment, place the dot before the debate, and again after.

Stress-Testing the Thesis

Six disciplinary lenses applied in parallel. Each row names how the lens challenges the central thesis (that play is the cognitive precondition for staying at the edge of chaos) and what the challenge implies for design or practice. At least one lens genuinely stresses the thesis, not just affirms it.

Six lenses on the central thesis
Lens	Challenge to the thesis	Implication
Cognitive & Behavioural Science	The thesis assumes adults retain the capacity to perform the cognitive operation de Mestral performed with the burr-and-microscope. The empirical record is mixed. Functional fixedness, expertise effects, and the well-documented decline of divergent-thinking scores with formal training all suggest that adult cognition is not merely playful cognition with a memo on top. The grammar may genuinely be different.	Design playful exercises as scaffolds for adult cognition (structured constraints, time-boxes, role inversions) rather than as invitations to “think like a child.”
Organizational Sociology	Even firms that adopt the language of play often sequester it in innovation labs and Greenhouse spaces while their core production system continues to reward predictability. The lab-versus-line split makes play visible without making it consequential. The thesis under-weights how completely incentive systems neutralise cultural overlays.	Couple every playful ritual with a corresponding change in performance evaluation; the play has to count toward someone's review for it to escape the lab.
Complex Adaptive Systems	The edge-of-chaos metaphor describes a system property emerging from local interactions, not a destination a strategist navigates to. A firm that explicitly tries to “locate itself on the edge of chaos” may produce the opposite: a centrally orchestrated proxy for adaptive behaviour that is itself the most ossified element of the operating model.	Build playful infrastructure as enabling constraint, not as choreographed performance. Make the edge of chaos the inevitable outcome of distributed permissions, not a managed posture.
Decision Theory & Biases	The Anti-Company Game and the Pre-Mortem assume that producing alternative scenarios disrupts the anchor of the current model. Empirically, generating alternatives often strengthens the anchor by making it the implicit reference point for every comparison. The exercise can systematise overconfidence in the present even as it appears to challenge it.	Pair each playful exercise with an explicit comparative protocol: score the alternatives against the baseline blind, with a third party adjudicating, to prevent the baseline from becoming the implicit winner.
Power & Political Economy	Playful exercises require executives to suspend the assumption that the current model is correct, which is also the assumption that justifies their seniority. The senior team that runs the Anti-Company Game is structurally the team least incentivised to take its conclusions seriously. The thesis under-weights how completely a firm's distribution of power resists its own re-imagination.	Run playful exercises with at least one external participant whose authority does not depend on the current model (a board director, an academic, a customer) so the conclusions have a sponsor who can survive them.
Information Theory	The Maverick Game treats small odd competitors as signal-rich. But for every Brooks Automation pivot there are hundreds of mavericks whose models genuinely do not work. The thesis under-specifies the discrimination problem: how to tell, in the moment, which maverick is a 1% prophet and which is 99% noise.	Track each maverick on a small set of structural indicators (unit economics direction, customer concentration trajectory, technology stack independence) rather than evaluating each on a curated narrative; the indicators discriminate where the narratives do not.

What survives: Play remains the cognitive precondition the thesis names, but institutionalising it requires structural pairing with incentive redesign, enabling constraint, explicit comparative protocols, and external sponsorship, none of which the soft cultural overlay it is usually marketed as can provide.

Questions This Paper Should Prompt

Where in your firm has the burr-under-the-microscope cognitive operation, the willingness to look twice at an observation everyone else has already brushed off, been retired into early management, and what would it take to bring it back into the strategy room?
If your annual partner retreat opened with a placement of your firm's dot on the edge-of-chaos 2×2, where would it sit today, and where would it have sat three years ago?
Which of your major clients is, right now, being silently outflanked by a maverick whose business model their management cannot explain, and is that maverick on your engagement file?
What is the basis-point cost of a deliberate play-capacity line in your practice budget, and what is the basis-point cost of not having one?

References

[1]
McGrath, R. G. (2013). The End of Competitive Advantage: How to Keep Your Strategy Moving as Fast as Your Business. Harvard Business Review Press. (Source for the “collapsing half-life” thesis and Figure 1 bar series.)
[2]
Daepp, M. I. G., Hamilton, M. J., West, G. B., & Bettencourt, L. M. A. (2015). The mortality of companies. Journal of the Royal Society Interface, 12(106): 20150120. (Constant-hazard, ~10-year half-life across 25,000 publicly traded firms, 1950–2009.)
[3]
Viguerie, S. P., Calder, N., & Hindo, B. (2021). 2021 Corporate Longevity Forecast. Innosight. (S&P 500 average tenure: ~33 years in 1965 → 15–20 years by the 2020s; projected 12–15 years by the early 2030s.)
[4]
Reeves, M., & Fuller, J. (2021). The Imagination Machine: How to Spark New Ideas and Create Your Company's Future. Harvard Business Review Press. (Origin work for the Anti-Company, Maverick, and Pre-Mortem game framework — Figure 4.)
[5]
Reeves, M., & Goodson, B. (2021). The Power of Anomaly. Harvard Business Review, July–August.
[6]
Klein, G. (2007). Performing a Project Premortem. Harvard Business Review, September.
[7]
Anthony, S. D. (2017). Anomalies Wanted. Forum for Growth and Innovation, Harvard Business School. (Christensen's anomaly-driven approach to refining theories of disruption.)
[8]
Gavetti, G., & Rivkin, J. W. (2005). How Strategists Really Think: Tapping the Power of Analogy. Harvard Business Review, 83(4), 54–63.
[9]
Wack, P. (1985). Scenarios: Uncharted Waters Ahead. Harvard Business Review, September–October. (Foundational Shell scenario-planning work.)
[10]
Herman, M., Frost, M., & Kurz, R. (2008). Wargaming for Leaders: Strategic Decision Making from the Battlefield to the Boardroom. McGraw-Hill.
[11]
Rehm, W., & Scanlan, J. (2017). How CFOs can use war gaming to support strategic decisions. McKinsey Quarterly, February.
[12]
Johnson, S. (2010). Where Good Ideas Come From: The Natural History of Innovation. Riverhead Books. (Adjacent-possible concept developed for strategic-imagination contexts; underlying biology from Kauffman.)
[13]
March, J. G. (1991). Exploration and Exploitation in Organizational Learning. Organization Science, 2(1), 71–87.
[14]
Edmondson, A. C. (1999). Psychological Safety and Learning Behavior in Work Teams. Administrative Science Quarterly, 44(2), 350–383. See also Edmondson, A. C. (2018). The Fearless Organization. Wiley.
[15]
O'Reilly, C. A., & Tushman, M. L. (2004). The Ambidextrous Organization. Harvard Business Review, April.
[16]
McGrath, R. G., & MacMillan, I. C. (1995). Discovery-Driven Planning. Harvard Business Review, July–August. (Real-options approach to strategic uncertainty.)
[17]
Peirce, C. S. (1931–1958). Collected Papers of Charles Sanders Peirce, 8 vols. Harvard University Press. (See especially Vol. 5, ¶171 and ¶189 on the deduction/induction/abduction trichotomy.)
[18]
Ribeiro, J., & Carmona, F. (2025). Analogical mappings of facts and counterfactuals in the human mind and Peirce's abduction: limitations in LLMs. Cognitive Systems Research / ScienceDirect (DOI: 10.1016/j.cogsys.2025.101426). (LLM limitations on counterfactual reasoning and true abductive inference.)
[19]
Liu, Y. et al. (2025). Hypothesis discovery in LLMs: a survey under Peirce's framework. Transactions on Machine Learning Research, April 2025.
[20]
Yamamoto, K. et al. (2026). Abductive Reasoning with Syllogistic Forms in Large Language Models. arXiv: 2603.06428.
[21]
Lipton, P. (2004). Inference to the Best Explanation (2nd ed.). Routledge. (The “lovely vs likely” distinction central to the Agentic-AI argument in Section 07.)
Supplementary & Parable Sources
•
Kauffman, S. A. (1995). At Home in the Universe: The Search for the Laws of Self-Organization and Complexity. Oxford University Press.
•
West, G. (2017). Scale: The Universal Laws of Growth, Innovation, Sustainability, and the Pace of Life. Penguin Press.
•
Catmull, E., with Wallace, A. (2014). Creativity, Inc.: Overcoming the Unseen Forces That Stand in the Way of True Inspiration. Random House. (Pixar Brain Trust source.)
•
Robertson, D., & Breen, B. (2013). Brick by Brick: How LEGO Rewrote the Rules of Innovation and Conquered the Global Toy Industry. Crown Business.
•
Sheff, D. (1993). Game Over: How Nintendo Conquered the World. Vintage.
•
Christensen, C. M. (1997). The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business School Press.
•
Cohen, M. D., March, J. G., & Olsen, J. P. (1972). A Garbage Can Model of Organizational Choice. Administrative Science Quarterly, 17(1), 1–25.
•
Reeves, M., Levin, S., & Ueda, D. (2016). The Biology of Corporate Survival. Harvard Business Review, January–February. (Systems-biology approach to corporate resilience.)
•
Page, S. E. (2007). The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies. Princeton University Press.

01The Bridge — Why This Question Is Urgent Now

02The Half-Life of Advantage and the Edge of Chaos

03Ten Threaded Parables

04The Grammar of Play — Eight Strategic Position Games

The Eight as a Sequence

05The Playgrounds — Environments Leadership Must Engineer

1. Psychological Safety

2. Slack — Structurally Protected Time

3. Diversity Within the Room

4. Tempo — Time-Boxed, Not Open-Ended

5. The Galápagos Principle — Distance From the Core

06Reading the Outcomes — From Play to Deployment

Reading Discipline 1 — Distinguishing Signal from Noise

Reading Discipline 2 — Real-Options Capital Allocation

Reading Discipline 3 — The Portfolio Logic

From Play to Deployment — The Convergence Move

07The Agentic AI Reframe — Why the Playground Becomes Existential

The Three Modes of Reasoning

Where Frontier AI Wins, and Where It Does Not

The Strategic Consequence

Three Implications for the Firm

08Playfulness for the Consulting Firm

Strategy — The Anti-Company Game in the Room

EPM Implementation on SAP and Oracle — The Playful Blueprint

09Resolution — The Threads Pull Together

1. The LEGO Group in 2003 — The Brick That Almost Vanished

2. The Pixar Brain Trust — The Studio and the Trust

3. The Nintendo Move of 1977 — The 134-Year-Old Toy Company

4. Fujifilm's Move — The Chemistry That Refused to Die With Its Market

5. Netflix's Move — The Firm That Killed Its Own Most Profitable Business

6. Microsoft's Resolution — The Company That Stopped Believing in Windows

7. Asian Paints' Counterintuitive Case — The Paint Company That Knew What Colour Your Living Room Wanted

8. Adobe's Move — The Shelf-Software Company That Walked Into the Cloud

9. Marvel's Move — The Comic Publisher That Walked Out of Bankruptcy and Into a Cinematic Universe

10. The Tata Group's Move — The In-House IT Department That Became a Global Services Champion

Author's Lens

Key Insights

Half-Lives Have Made Re-Imagination the Primitive Move

Play Is the Cognitive Precondition, Not the Cultural Flourish

The Edge of Chaos Is a Strategic Position, Not a Personality

The Most Dangerous Spreadsheet Is the One That Was Right Last Quarter

Consulting Is the Practice Most Exposed to Its Own Diagnosis

Play Is Procurable, but Only If It Becomes a Line Item

Actionable Takeaways

Stress-Testing the Thesis

Questions This Paper Should Prompt

References

Intelligent and Adaptive Finance

Quick Links

Stay Connected