The Patterns

Ten ways a good fix backfires. Every case we collect is taken apart and tagged with the patterns behind it, so the next builder can spot one before shipping it.

  • 10patterns
  • 16cases tagged
  • 30research sources

Start with pattern 1Open the dashboards

All ten patterns

  1. 01Solve a shortage, create a surplus Too little becomes too much, and the extra goes somewhere. Ha
  2. 02Too stable to go away Built to last, so it lasts where it shouldn't. BaMüMiTh
  3. 03The problem moves Fixed in one place, it lands in another. NoMiBeHaBaMoJuThZu
  4. 04Great for one, bad for all Each user wins. When everyone uses it, everyone loses. MoThZuBaJuBo
  5. 05Chasing the wrong number It hit its target. The target was the wrong one. Zu
  6. 06The cure breeds resistance Kill most of the problem and the survivors take over. JuFrMü
  7. 07The reward gets gamed People chase the incentive instead of the goal. Zu
  8. 08Cheap now, paid later The bill arrives decades after the benefit. BaBoBaSi
  9. 09The lesson flipped Built to teach one lesson, it taught the opposite. Ma
  10. 10The risk piles up Stop every small failure, and the risk saves up for one big one. Si

Solve a shortage, create a surplus

Too little becomes too much, and the extra goes somewhere.

The fix raises the supply of something scarce so well that the excess, beyond what the system can absorb, builds up and does harm.

Ask before you ship

  1. If this works better than planned, where does the extra go?
  2. What absorbs the excess today, and how much more can it take?

It counts when

  • The original problem was a shortage.
  • The fix scaled up supply.
  • The harm comes from the excess, not from the thing in normal amounts.

Easy to confuse with

The problem moves, when the excess lands somewhere else. Tag both when it does.

Cases

In the research

3 sources

Nutrient surplusOverproduction under price supportsRebound effect (Jevons paradox)Overshoot

  • Meadows, Meadows, Randers and Behrens, The Limits to Growth (1972)
  • Erisman and others, "How a century of ammonia synthesis changed the world," Nature Geoscience (2008)
  • Jevons, The Coal Question (1865)

Too stable to go away

Built to last, so it lasts where it shouldn't.

The property that made the fix useful, its durability, keeps it in the world long after its use, so it builds up or travels to where it does harm.

Ask before you ship

  1. How long does this last after we stop needing it?
  2. Where does it end up, and what breaks it down there?

It counts when

  • Its usefulness depended on lasting.
  • The harm comes from lasting past its use or moving beyond it.

Easy to confuse with

Cheap now, paid later, which is about when the bill arrives, not why the thing persists.

Cases

In the research

2 sources

PersistenceBioaccumulation

  • Stockholm Convention on Persistent Organic Pollutants, Annex D screening criteria (2001)
  • Geyer, Jambeck and Law, "Production, use, and fate of all plastics ever made," Science Advances (2017)

The problem moves

Fixed in one place, it lands in another.

The fix solves the problem where it is measured and shifts the cost to a place, group or system the decision makers weren't counting.

Ask before you ship

  1. Who isn't at the table who will feel this?
  2. Where does the problem go after it leaves our boundary?

It counts when

  • The harm lands outside the boundary the fix was judged within.
  • The people or places harmed bore the cost without being counted in the decision or compensated for it.

Easy to confuse with

Cheap now, paid later, when the same users pay later instead of someone else.

Cases

In the research

2 sources

ExternalityProblem displacementRearranging effects

  • Pigou, The Economics of Welfare (1920)
  • Tenner, Why Things Bite Back (1996)

Great for one, bad for all

Each user wins. When everyone uses it, everyone loses.

Each adopter gains from the fix, but adoption draws down or crowds a shared resource no one can be kept from using, so when everyone adopts, the shared result is worse than before.

Ask before you ship

  1. What happens when everyone does this, not just us?
  2. What shared thing gets used up when it spreads?

It counts when

  • Adopting is the sensible choice for each person or firm.
  • The harm comes from the total, not from any one user.

Easy to confuse with

The problem moves, when the harm falls on outsiders rather than on the users as a group. Induced demand belongs here when new capacity fills up because everyone shifts to it.

Cases

In the research

4 sources

Tragedy of the commonsSocial trapInduced demand

  • Hardin, "The Tragedy of the Commons," Science (1968)
  • Platt, "Social Traps," American Psychologist (1973)
  • Downs, "The Law of Peak-Hour Expressway Congestion," Traffic Quarterly (1962)
  • Ostrom, Governing the Commons (1990), on how communities avoid the tragedy

Chasing the wrong number

It hit its target. The target was the wrong one.

The fix improved a measurable stand-in for the goal, but the stand-in was not on the causal path to the goal, so it rose while the goal didn't, or got worse.

Ask before you ship

  1. If this number goes up and nothing else changes, are we better off?
  2. What could move this number without helping anyone?

It counts when

  • The designers chose a measure to optimize.
  • The measure and the real goal came apart.

Easy to confuse with

The reward gets gamed, where people exploit the measure. Here the measure itself was the wrong target. Tag both when both happen.

Cases

In the research

2 sources

Surrogate endpointSurrogationProxy failure

  • Fleming and DeMets, "Surrogate end points in clinical trials: are we being misled?" Annals of Internal Medicine (1996)
  • Choi, Hecht and Tayler, "Lost in translation: The effects of incentive compensation on strategy surrogation," The Accounting Review (2012)

The cure breeds resistance

Kill most of the problem and the survivors take over.

The fix suppresses most of a living population, which selects for heritable traits that let some survive it, until the survivors are the problem.

Ask before you ship

  1. What survives this, and what happens when that's all that's left?
  2. Are we relying on one control?

It counts when

  • The target population has heritable variation and reproduces fast enough for selection to act.
  • The fix was applied widely or alone, with few other controls.

Easy to confuse with

The reward gets gamed, which is people adapting to incentives, not populations evolving.

Cases

In the research

3 sources

Evolved resistanceSelection pressureAntimicrobial resistance

  • Fleming, "Penicillin," Nobel Lecture (December 11, 1945)
  • Antimicrobial Resistance Collaborators, "Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis," The Lancet (2022)
  • Heap, The International Herbicide-Resistant Weed Database, weedscience.org (accessed October 2026)

The reward gets gamed

People chase the incentive instead of the goal.

The fix created a reward, and people found ways to earn it that don't serve, or that work against, the goal it was meant to reach.

Ask before you ship

  1. What's the cheapest way to earn this reward?
  2. Would we be happy if people did exactly that?

It counts when

  • There is an incentive: money, status, a score or a rule.
  • People change their behavior to earn it.

Easy to confuse with

Chasing the wrong number, where the designers picked the wrong target before anyone gamed it.

Cases

In the research

3 sources

Goodhart's lawCampbell's lawPerverse incentive

  • Kerr, "On the Folly of Rewarding A, While Hoping for B," Academy of Management Journal (1975)
  • Goodhart, "Problems of Monetary Management: The U.K. Experience," Papers in Monetary Economics, Reserve Bank of Australia (1975)
  • Campbell, "Assessing the impact of planned social change," Evaluation and Program Planning (1979)

Cheap now, paid later

The bill arrives decades after the benefit.

The benefit arrives at once and the cost arrives years or generations later, so it barely weighs on the decision.

Ask before you ship

  1. What does this cost in twenty years, and who pays then?
  2. What would we decide if the bill came due today?

It counts when

  • A long delay separates the benefit from the harm.
  • The delay is part of why the harm was discounted or missed.

Easy to confuse with

Too stable to go away, which explains why some harms last, not why they were discounted.

Cases

In the research

3 sources

Delayed feedbackTemporal discounting

  • Senge, The Fifth Discipline (1990)
  • Sterman, Business Dynamics (2000)
  • Frederick, Loewenstein and O'Donoghue, "Time Discounting and Time Preference: A Critical Review," Journal of Economic Literature (2002)

The lesson flipped

Built to teach one lesson, it taught the opposite.

Something made to teach, warn or persuade ends up teaching or encouraging the opposite of its message.

Ask before you ship

  1. Could the audience take the opposite message from this?
  2. What will people enjoy about the thing we're warning against?

It counts when

  • The fix had a message or a lesson.
  • The audience took away the reverse.

Easy to confuse with

The reward gets gamed, when the reversal comes from an incentive rather than from the message.

Cases

In the research

4 sources

Boomerang effectPsychological reactanceDescriptive norm backfire

  • Brehm, A Theory of Psychological Reactance (1966)
  • Cialdini, "Crafting normative messages to protect the environment," Current Directions in Psychological Science (2003)
  • Hornik and others, "Effects of the National Youth Anti-Drug Media Campaign on Youths," American Journal of Public Health (2008)
  • Byrne and Hart, "The boomerang effect: A synthesis of findings and a preliminary theoretical framework," Communication Yearbook (2009)

The risk piles up

Stop every small failure, and the risk saves up for one big one.

The fix prevents small, frequent failures so well that the risk they used to release builds up until a rare, much larger failure.

Ask before you ship

  1. What small failures are we preventing, and what were they releasing?
  2. What happens on the day the protection fails?

It counts when

  • The fix suppresses small failures.
  • Something accumulates behind it: fuel, development in harm's way, debt or deferred losses.

Easy to confuse with

Cheap now, paid later, which is about delay. Here the harm is held back and then released all at once.

Cases

In the research

4 sources

Levee effectSafe development paradoxFire suppression paradoxMoral hazard

  • White, Human Adjustment to Floods (1945)
  • Burby, "Hurricane Katrina and the Paradoxes of Government Disaster Policy: Bringing About Wise Governmental Decisions for Hazardous Areas," The Annals of the American Academy of Political and Social Science (2006)
  • Calkin, Cohen, Finney and Thompson, "How risk management can prevent future wildfire disasters in the wildland-urban interface," PNAS (2014)
  • Pauly, "The Economics of Moral Hazard: Comment," American Economic Review (1968)

What 16 backfires have in common

The patterns, the scale, the time lag and who paid, counted across every case.

Open the dashboards

Other tags on every case

Scale

How far the harm reached.

LocalRegionalNationalGlobal

Time lag

How long from the fix to the harm being felt.

Under a yearYearsDecadesGenerations

Who pays

Who carried the cost. Can be more than one.

The people it servedBystanders and other groupsFuture generationsEcosystems and other species

Reversibility

How hard the harm is to undo.

ReversibleSlow to undoIrreversible

Tags are provisional. Two separate automated passes coded every case and an editor settled the differences. Human coders review them next, and the codebook gets a new version number whenever a definition changes.