Finagle's Law

Every operations leader has a story about the outage that hit during the product launch, or the key supplier that went dark the week before the board meeting. Engineers gave that pattern a name decades ago: Finagle's Law, the wry cousin of Murphy's Law that adds a cruel refinement — failures do not just happen, they happen at the worst possible moment. The saying began as shop-floor humor, but it describes a real property of complex systems under stress. Executives who take it seriously build organizations that bend under pressure instead of breaking, while those who dismiss it as a joke tend to relearn the lesson during an actual crisis.

The origin of a useful joke

The phrase traces to Larry Niven's 1970s science fiction, where a character invokes "Finagle's Law of Dynamic Negatives" as a darker sibling of the better-known Murphy's Law1. Murphy's Law itself grew out of 1949 rocket-sled tests at Edwards Air Force Base, where an engineer's frustration with a miswired sensor turned into the now-familiar adage that anything that can go wrong will go wrong. Niven's addition sharpened the idea by attaching a timing clause: the failure surfaces not randomly but precisely when it does the most damage. That distinction matters more than it sounds. A random failure is a nuisance. A failure timed to your weakest moment, whether that is peak season, a system migration, or a leadership transition, is a threat multiplier. Engineers who spent careers building redundant systems understood this instinctively long before the phrase had a name. Business leaders are still catching up.

Why optimism hides the exposure

Most planning processes reward confident forecasts, not honest risk disclosure. A project sponsor who flags every way a plan could unravel looks unhelpful next to one who promises a clean rollout and that incentive structure quietly builds fragility into the organization. Teams anchor on the most likely scenario and underweight the scenario where two problems compound at once, which is exactly the scenario Finagle's Law predicts will occur. The result is plans that work in the boardroom rehearsal and fail in the field. Behavioral research describes this pattern as the planning fallacy, the tendency to estimate a task's timeline and cost from a best-case scenario rather than from how similar projects have actually gone and it shows up across domains from construction schedules to merger integrations2. Consulting engagements regularly uncover the same gap between the plan a client presented and the plan a client actually followed. The fix is not pessimism for its own sake. It is a structured way to force the uncomfortable scenario onto the table before it forces itself onto the calendar.

Making bad news travel fast

One structured technique with strong staying power is the premortem, developed by psychologist Gary Klein and popularized through Harvard Business Review. A team is told to imagine the project has already failed, then works backward to write down every plausible reason why, before the project starts rather than after it stumbles3. The exercise works because it removes the social cost of dissent. Nobody is predicting failure; everybody is explaining a failure that has already, hypothetically, happened. That framing pulls out risks that a standard risk register misses, particularly the ones tied to timing and coincidence rather than a single obvious cause. Leaders who run a premortem before a major initiative typically walk away with a shorter, sharper list of mitigations than the one their original risk assessment produced. The exercise takes an afternoon. Skipping it can cost a quarter.

By making it safe for dissenters who are knowledgeable about the undertaking and worried about its weaknesses to speak up, you can improve a project's chances of success

That is the social dynamic Klein's premortem was built to defeat and it is the same dynamic Finagle's Law exploits in any organization that punishes pessimism. Nobody wants to be the person who predicted the outage that then, months later, actually happens on schedule.

Removing single points of failure

Resilience engineering offers the clearest antidote to Finagle's Law: assume the failure will land at the worst time, then design so that no single failure can take the whole system down with it. That means redundant suppliers instead of one preferred vendor, cross-trained staff instead of one irreplaceable expert and decision rights that do not collapse if one executive is on a plane. McKinsey's research on organizational resilience frames this as a shift from optimizing purely for efficiency toward deliberately carrying some slack, arguing that the firms best positioned to handle shocks are the ones that built buffer capacity before they needed it4. That tradeoff is real and worth naming honestly: redundancy costs money and slows some processes down. The alternative cost, paid only when Finagle's Law collects, is usually higher. Boards that treat resilience spending as a discretionary line item are underpricing the risk they are carrying.

Testing the plan before the crisis does

A continuity plan that has never been tested is a set of assumptions with a nice cover page. Financial institutions and other regulated firms have learned this the hard way, which is why business continuity planning now typically includes scheduled failover drills, tabletop exercises and after-action reviews rather than a document that sits untouched until an incident forces everyone to open it for the first time5. The value of the drill is not the document it produces. It is the gap the drill exposes between what the plan assumes and what the organization can actually do under pressure, on a Friday evening, with half the usual staff available. Companies that run these exercises regularly tend to find smaller, cheaper problems each time, because the big ones got fixed the first year. Companies that skip the drill find the big problems during the real event, when fixing them costs far more than a training afternoon.

Building a culture that surfaces problems early

Structural fixes only work if the people closest to a problem feel safe reporting it before it becomes a crisis. That requires deliberate cultural work: leaders who ask "what's not working" as often as they ask "what's the status", escalation paths short enough that a junior engineer can flag an anomaly without three layers of approval and a track record of treating early warnings as useful rather than as evidence of failure. Teams that hide problems until they are unavoidable guarantee that those problems will be discovered under the worst possible conditions, which is Finagle's Law operating exactly as advertised. Workplaces where people report feeling safe to flag mistakes tend to catch small problems before they compound into large ones, a pattern documented across studies of high-performing teams6. The organizations that handle disruption well are usually the ones where a mid-level manager felt comfortable raising a concern two weeks before it mattered. That comfort is a leadership choice, reinforced or destroyed one reaction at a time. It costs nothing to build and very little to destroy, which is why it deserves deliberate attention rather than an assumption that it will take care of itself.

Sizing the resilience investment correctly

Not every risk deserves the same countermeasure and treating all failure modes as equally worthy of redundancy spreads resources thin without reducing real exposure. The more useful exercise is ranking failure scenarios by two factors: how likely the failure is and how much worse it gets if it lands during a period of thin margin. A supplier delay during a slow month is an inconvenience. The same delay during peak season, layered on top of a system migration, is the scenario Finagle's Law is describing. Leaders who map their specific worst-timing scenarios, rather than buying generic insurance against generic risk, spend their resilience budget where it actually reduces exposure. That mapping exercise, done once a year and revisited after any major organizational change, keeps the resilience program tied to the business as it actually operates rather than to a static risk register from three years ago.

• 1Murphy's law
• 2Planning fallacy
• 3Performing a project premortem
• 4Building organizational resilience
• 5What is a business continuity plan
• 6Signs you work in a psychologically safe workplace