An anti-pattern is a recurring solution that looks plausible but backfires: it solves a surface problem while creating worse ones or locking in failure. The term came from software (design patterns that go wrong), but the mental model applies anywhere: organisational habits, strategy playbooks, and decision rules that reliably produce bad outcomes.
Recognising anti-patterns protects you twice. First, you avoid adopting them — when a "best practice" keeps failing in your context, you may be watching an anti-pattern. Second, you can name and escape them when you're already inside one. "We're doing a death march" or "this is analysis paralysis" gives the team a shared label and permission to change.
Classic examples: Golden hammer — using one tool for every problem. Scope creep — adding requirements until the project can't ship. Analysis paralysis — delaying decisions by demanding more data. Design by committee — diluting a clear vision into consensus mush. Premature optimisation — polishing the wrong thing. In each case, the pattern feels like the right move (standardise, be thorough, get alignment) but the outcome is worse.
The value is in the catalogue. Once you know the list, you can spot the shape of failure before it completes. You can also audit your own org: which of our standard responses are actually anti-patterns in this environment?
Section 2
How to See It
You see anti-patterns when the same type of solution keeps leading to the same type of failure. The diagnostic: repeated application of a "reasonable" approach, repeated bad result. When post-mortems keep citing the same cause ("we didn't have enough process" or "we had too much process"), you're likely in anti-pattern territory.
Business
You're seeing Anti-patterns when every product launch is delayed by last-minute scope adds. The pattern — "one more feature" — is consistent; the outcome is consistent (slipped dates, diluted focus). The solution that feels right (please stakeholders) is the one causing the failure.
Technology
You're seeing Anti-patterns when the team rewrites the system every few years chasing the latest stack. The pattern — "we need to modernise" — produces a new codebase that will itself be replaced. The anti-pattern is the rewrite loop, not any single technology choice.
Investing
You're seeing Anti-patterns when a fund keeps doubling down on the same thesis after repeated drawdowns. The pattern — "we're right, the market is wrong" — is a form of escalation. Naming it (irrational escalation, sunk cost) is the first step to changing behaviour.
Markets
You're seeing Anti-patterns when regulators respond to each crisis with more rules, and the next crisis emerges from a new gap. The pattern — "add regulation where the last failure was" — can create complexity and false security without fixing systemic risk.
Section 3
How to Use It
Decision filter
"When a standard approach keeps failing, ask: is this an anti-pattern? Name it. Then ask what would break the pattern — often the opposite move (less process, fewer stakeholders, smaller scope) or a different structure entirely."
As a founder
Audit your org for anti-patterns. Do you have design by committee? Scope creep? Hero culture (only one person can fix things)? Naming them gives you permission to change. The fix is usually structural: different decision rights, different incentives, or a deliberate counter-pattern (e.g. "no new scope after this date").
As an investor
Watch for anti-patterns in portfolio companies: rewrite obsession, analysis paralysis, or consensus-driven strategy. The pattern often persists because it feels professional or safe. Your job is to name it and push for the counter-move before the company pays the full cost again.
As a decision-maker
Before adopting a "best practice," check whether it has failed in similar contexts. If the same approach keeps failing elsewhere, it may be an anti-pattern in this domain. Prefer experiments over wholesale adoption until you've seen it work in your environment.
Common misapplication: Calling every failure an anti-pattern. The definition requires a recurring bad solution. One-off mistakes are just mistakes. Second misapplication: Assuming the opposite of an anti-pattern is always right. Sometimes the fix is a different structure, not the inverse (e.g. not "no process" but "process in the right place").
Grove was explicit about avoiding anti-patterns: "Only the paranoid survive" was a guard against complacency as a pattern. He institutionalised constructive confrontation to break the anti-pattern of consensus and unclear disagreement. Naming the danger (e.g. the "wrong business" or strategic inflection) was a way to force a different response.
Netflix's "no vacation policy" and "keeper test" were deliberate counter-patterns to bureaucracy and low accountability. Hastings has described avoiding the anti-pattern of process creep as the company scaled — adding rules when something went wrong often made things worse. The alternative was clarity of context and high talent density.
Section 6
Visual Explanation
Anti-patterns: a recurring solution that feels right but produces worse outcomes. Break the loop by naming it and changing structure or incentives.
Section 7
Connected Models
Anti-patterns connect to models about patterns, failure, and second-order effects. The grid below shows what reinforces the idea, what creates tension, and what it leads to.
Reinforces
Design Pattern
Design patterns are proven solutions to recurring problems; anti-patterns are proven bad solutions. The same discipline — name the recurrence, document the context — applies. Recognising one helps you recognise the other.
Reinforces
Second-Order Effects
Anti-patterns often succeed on first-order goals (e.g. alignment) and fail on second-order ones (speed, ownership). Thinking in second-order effects helps you spot when a "solution" is actually an anti-pattern.
Tension
Cargo Cults
Cargo cults copy form without function; anti-patterns are copied forms that actively backfire. The tension: both involve mimicry, but anti-patterns have a clearer causal link to repeated failure. Escaping cargo cults means finding the function; escaping anti-patterns means changing the structure.
Tension
Sunk [Cost](/mental-models/cost) Fallacy
Sunk cost keeps you in a bad course; anti-patterns are bad courses you keep repeating. The tension: stopping an anti-pattern may feel like admitting past waste. The move is to frame it as stopping future waste, not defending past decisions.
Leads-to
Section 8
One Key Quote
"Anti-patterns are the opposite of best practices — they are worst practices."
— Andrew Koenig (on anti-patterns in software)
The point is not that people are stupid; it's that plausible, even professional-looking approaches can be reliably worse. Naming them turns invisible repetition into something you can change.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Build a short list of anti-patterns for your domain. Scope creep, design by committee, analysis paralysis, golden hammer — know the shapes. When you see the same failure mode again, check if a named anti-pattern fits. If it does, you have a lever: change the structure or incentives that reinforce it.
The fix is usually structural. Adding another policy or review often is the anti-pattern. The real fix is decision rights, incentives, or constraints (e.g. hard scope freeze, single decision-maker). Ask what would make the anti-pattern impossible rather than asking people to "try harder."
Audit your own org. What do we do every time we have a problem? If the answer is "add process" or "get more alignment," and you've seen that fail before, you may be in an anti-pattern. Name it at the leadership level and design the counter-move.
Section 10
Test Yourself
Is this mental model at work here?
Scenario 1
After every incident, the company adds a new approval step. Projects slow down; incidents keep happening in unanticipated ways.
Scenario 2
A startup fails once because it ran out of cash. The founder says 'we need more runway next time.'
Scenario 3
Every quarter the board asks for more detailed forecasts. The team spends two weeks on slides. Forecasts are still wrong; the next quarter the board asks for even more detail.
Scenario 4
A company copies a competitor's org structure because 'it works for them.' Performance doesn't improve.
Section 11
Summary & Further Reading
Summary: Anti-patterns are recurring solutions that look reasonable but produce worse outcomes. Recognising them lets you avoid adopting them and escape them when you're in one. Name the pattern, then change the structure or incentives that keep it in place.
Not about anti-patterns per se, but about how design can create repeated failure modes. Useful for thinking about structural causes of bad outcomes.
Premature Optimisation
Premature optimisation is a specific anti-pattern: optimising the wrong thing. Naming it as an anti-pattern helps teams avoid it by default and focus on the bottleneck first.
Leads-to
Root Cause Analysis
When the same failure recurs, root cause analysis asks why. Anti-patterns are often the why — the repeated response that generates the failure. Fixing the root cause may mean eliminating the anti-pattern, not just fixing the immediate incident.
