How to Run a Pre-Mortem: Kill Your Strategy Before the Market Does

In December 1975, a 24-year-old Kodak engineer named Steve Sasson wired a CCD sensor into a toaster-sized contraption of spare parts and pointed it at a colleague.¹ Twenty-three seconds later, a blurry black-and-white image resolved onto a television screen — a photograph captured with no film at all. Sasson carried the prototype to Kodak's management. Their verdict, as he recounted it years afterward, was that the thing was a novelty not worth mentioning outside the building.³ Thirty-seven years later, Kodak filed for Chapter 11 bankruptcy.¹ The company wasn't destroyed by a technology it failed to see coming. It was destroyed by a technology it built, demonstrated to its own leadership, and chose never to interrogate.

That distinction is the entire argument for a pre-mortem. Most planning meetings run on institutional optimism — the people proposing a bet are usually the same people selling it, and asking "how could this fail?" out loud can read as disloyalty. A pre-mortem inverts the incentive. You assume the decision has already failed, put a real date on the failure, and get rewarded — for that one meeting — for explaining exactly why. It isn't a forecasting exercise; nobody has to predict the future correctly. It's a discipline: naming, in advance, the failure modes that are already visible to somebody in the room, before they show up on a balance sheet. Kodak had the prototype years before digital photography became a real consumer market. What it never had was that meeting.

Why protecting film felt like the only responsible choice: the recurring-revenue math that made a camera an existential threat, not a product line

By the mid-1970s, Kodak wasn't just a camera company; it owned nearly the entire chain photography ran through. At its peak in the mid-1990s, the company's market value topped $28 billion, built on a business that sold film, then sold the processing, then sold the paper and chemicals underneath every print.⁵ By 1976, Kodak held roughly 90% of the U.S. film market and 85% of camera sales.² Every one of those numbers depended on a customer coming back to buy more film. A digital camera has no consumable. Buy it once, and the recurring purchase — the actual engine of Kodak's profit — disappears entirely. It's the razor-and-blades model in reverse: Kodak sold the razor and the blades both, forever, and digital was a razor with no blade at all. Digital wasn't a new competitor to one Kodak product line. It was a design for eliminating the mechanism the whole business model ran on. Protecting film wasn't blindness. It was the locally correct read of that year's numbers, repeated for two decades running.

The method: write the failure before it happens: four steps, run before the decision, not after it's already gone wrong

Run a pre-mortem on any live decision and the mechanics don't change by industry. First, assume the decision has already failed and write the headline as fact, not risk — not "this could go wrong" but "this failed, and here is the article about it." Second, list every cause that could have produced that headline, including the ones that implicate the people in the room. Third, rate each cause's likelihood and name the earliest signal that would have given warning — the tripwire you'd have seen months before the outcome. Fourth, and hardest: isolate the single assumption that, if wrong, takes the entire plan down on its own, independent of every other cause on the list. That last step is the one Kodak's 1975 meeting never reached. The load-bearing assumption — that customers would keep paying for the film-and-processing loop even after a filmless alternative existed — was never written down as a bet that could lose.

The part that actually breaks: saying the uncomfortable cause out loud: a pre-mortem only works if someone in the room is willing to name the cause that indicts the people running it

A pre-mortem isn't hard technically. It's hard politically. The failure causes worth naming are rarely about the market or a competitor's timing — they're about your own incentives, and those are the causes a room full of people paid by the existing business will not volunteer first. That's an uncomfortable thing to ask a room to do, and it's precisely why most pre-mortems quietly default to the safe causes — bad luck, timing, a competitor nobody could have predicted — and skip the one that would require someone to admit the plan itself, or the person championing it, is the problem. Kodak's management had, by any fair reading, already run half a pre-mortem: they had seen the prototype and understood exactly what it implied for film. What never happened was the second half — naming, out loud, in a room with the authority to act, that the assumption underneath the entire film business was on borrowed time. Sasson recorded their answer instead of the analysis:

That isn't a failure of forecasting. Kodak's own market-intelligence team ran the numbers in 1981 and correctly concluded the company had roughly a decade before digital became a real threat to film⁷ — a genuine pre-mortem, just one nobody acted on. What Kodak lacked wasn't the analysis. It was the nerve to let an uncomfortable conclusion change what the company actually did. Running the method doesn't require a genius forecaster. It requires someone in the room willing to act on what the forecast already says.

The control case: the identical threat, the opposite pre-mortem: Fujifilm faced the same collapse and answered the load-bearing question differently

Fujifilm is the cleanest natural experiment strategy has to offer: same industry, same collapsing product, same decade, opposite outcome. When Shigetaka Komori became Fujifilm's CEO in 2003,⁶ he inherited a company facing the identical extinction event Kodak's did — the same collapsing film market, the same commodity-priced digital cameras arriving to replace it. Komori's version of the load-bearing assumption ran the opposite direction from Kodak's. Instead of asking how to keep selling film a little longer, he asked what Fujifilm actually knew how to do that didn't depend on film surviving at all. The answer had been sitting underneath the product the whole time: the collagen and antioxidant chemistry Fujifilm had spent decades refining to keep prints from fading was, at the molecular level, skin-care science. Fujifilm redeployed it into a cosmetics line, Astalift, and pushed the same underlying expertise into healthcare and advanced materials.⁶ In 2012 — the same year Kodak filed for bankruptcy — Fujifilm posted record revenue above $21 billion.⁶ Same collapsing market. Same warning signs. A load-bearing assumption tested instead of protected.

The honest objection is that Kodak's dismissal looks worse in hindsight than it felt in the room. Even a perfect pre-mortem doesn't guarantee a company can win the business it's warning itself about. Digital cameras became a thin-margin hardware category that even the winners struggled to hold once smartphones absorbed it entirely, so running the method wouldn't necessarily have handed Kodak a second version of its film business — the honest answer to "what happens if digital wins" may genuinely have been "there's no equally good business on the other side of this." That doesn't excuse the choice Kodak actually made, which was never to write the risk down as a real bet at all. But it's worth admitting that a good pre-mortem can identify a real threat without producing a comfortable answer. Naming the load-bearing assumption is necessary. It is never sufficient by itself — someone still has to be willing to act on what it finds, the way Fujifilm's leadership had to overrule its own instinct to keep defending film a little longer.

The method doesn't require predicting the future. Sasson's camera proved Kodak didn't need to — its own analysts got the timeline right in 1981. What it requires is writing down, on purpose and in advance, what a room full of smart people already half-know, and then being willing to read that sentence back out loud before the failure, not after it. Run it on the plan you're funding this quarter, not the one you'll be explaining next year.