The Reverse Anna Karenina Principle
Dysfunctional companies all fail the same way; high performers succeed in wildly different ways.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 4
- Confidence
- 93%
Tolstoy said happy families are all alike and unhappy families are each unhappy in their own way. In product organisations, John Cutler argues the reverse holds: the dysfunction is stereotyped and recognisable, while high performance is achieved through radically different structures, cultures and decision styles. The practical consequence is that you can diagnose reliably from anti-patterns, but you cannot copy a high performer's mechanism and expect it to transfer. Improvement means removing your specific anti-patterns and then finding a coherent path that matches who you actually are.
Origin
Cutler credits the principle to his friend Josh Arnold, who coined it as an inversion of the Anna Karenina principle from Leo Tolstoy's novel. Cutler says it held true across the hundreds of product teams he worked with as Amplitude's product evangelist.
Core principles
- 01Anti-patterns are near-universal and easy to name; excellence is context-specific and hard to name.
- 02High-level success principles ('make good decisions faster', 'trust each other') are true but non-transferable as tactics.
- 03The same outcome — e.g. fast, good decisions — is reachable by rigorous process, by serendipitous collaboration, or by a top-down CEO who simply decides.
- 04Copying another company's mechanism without its context often does more harm than doing nothing.
How to run it
- 1
Diagnose from the anti-pattern side, not the excellence side
List the failure patterns you can actually see in your org — feature-factory shipping, strategy-structure mismatch, no learning loop, incoherent leadership. These are the reliable signal, because they repeat across every dysfunctional company.
Pro tip If a description of an anti-pattern makes people say 'how did you know that was happening in my company?', you have found the shared failure mode, not a company-specific one.
- 2
Name the high-level outcome you want, not the mechanism
State the outcome abstractly: 'we make better decisions, faster.' Then decompose what any route to that outcome requires — information, diverse perspectives, analytical capability, domain chops, and a goal you are optimising for.
Watch out Stopping at the abstract principle is unsatisfying but necessary — jumping straight to a mechanism is where imported frameworks go wrong.
- 3
Enumerate at least three legitimate mechanisms that reach it
For the decision-quality example: (a) a rigorous process-driven company that red-teams its decisions, (b) a loosely-structured company that relies on serendipitous connections between the right people, (c) a top-down company where the CEO decides the roadmap and hires people who are happy with that. All three exist and all three succeed.
Pro tip Forcing yourself to name a third and fourth mechanism breaks the assumption that your preferred one is the only one.
- 4
Pick the mechanism that is coherent with who you and your people actually are
Choose the route your leadership and culture can actually sustain, not the one that has the best press. A process-driven meritocratic leader cannot become a consensus-driven collectivist leader; they can only flex slightly from their happy place.
Pro tip Test coherence by asking whether your stated values and your observable behaviours match. Empowerment talk plus command-and-control behaviour is incoherence, and it is visible to everyone.
Watch out Do not use 'we're just different' as cover for leaving anti-patterns in place. Step 1 is not optional.
In the wild
Cutler describes idealistic product managers insisting decision-making must be pushed down to teams, while he watches a company where the CEO simply tells everyone what to do — and it works, because that company attracted people who don't mind it and who buy into the vision.
→ The company performs well despite violating a canonical product principle, showing that the mechanism, not the outcome, is what varies.
Cutler repeatedly met teams that had hired brilliant people yet were struggling because funding, incentives, org structure and technical architecture did not line up with the current strategy.
→ No amount of empowerment or talent density unblocked them; the fix required realigning structure to strategy, which is an anti-pattern removal, not an excellence import.
Common mistakes
Treating another company's practices as a recipe
Reading how Figma or Amazon works and installing their mechanisms ignores that their mechanism is coherent with their context, not yours. Cutler saw tools deployed out of context produce worse effects than not deploying them at all.
Assuming there is one true way to be high-performing
Humble servant leaders and dominant sparring leaders both build great teams. Ranking companies on a single high-performing/low-performing spectrum hides the diversity of legitimate paths.
Is it for you?
Best for
Product leaders and founders benchmarking themselves against 'best practice' content, or consultants diagnosing an unfamiliar org.
Not ideal for
Teams that have no obvious anti-patterns left and need a positive, concrete next play — the principle diagnoses but does not prescribe.
From the transcript
“the dysfunctional families are all different and the happy families are the same and what he said is it's the reverse of that with product…”
“either you have the very easy to identify anti-patterns or you have the high level principles”
“you can have brilliant teams you meet these brilliant teams where they've just hired in like the quote unquote Best of the Best and they're…”
From the episode
What differentiates the highest-performing product teams
John Cutler (Amplitude, The Beautiful Mess)