Core Values for Innovation: The Permission Structure

The Vision/Traction Organizer opens with core values, and this is the section serious people skim — because most core values are the poster in the lobby, and everyone has seen the poster fail. The canonical failure has a name and a court record: Enron’s stated values, printed in its 2000 annual report and a sixty-four-page Code of Ethics now preserved in the Department of Justice’s archives, were Communication, Respect, Integrity, and Excellence — the code even declared that “ruthlessness, callousness and arrogance don’t belong here.” Patrick Lencioni opened his 2002 HBR piece “Make Your Values Mean Something” with exactly this corpse, and organizational theory had already named the disease decades earlier: Argyris and Schön’s distinction between espoused theory (what we say we believe) and theory-in-use (what our behavior reveals we believe). A values statement is an espoused theory. The org chart of who got promoted last year is the theory-in-use.

So why does Traction put values first, and why should an innovation strategist — of all people — care? Because innovation runs on something operations doesn’t: permission that can’t be requested. Nobody asks “may I have an idea that threatens our main product?” in a meeting. Whether that idea gets voiced, prototyped, or quietly strangled in the shower is decided by each person’s private prediction of what happens to people who try — and that prediction is your real values system, the theory-in-use. Wickman’s practical rule — three to seven values, actually used to hire, review, reward, and fire — is the mechanism that closes the espoused/enacted gap: a value you’ve never fired or promoted against is a decoration. This post is about which values, specifically, an innovation strategy needs enacted, and what the good versions look like on paper.

The 1948 original: McKnight’s tolerance clause

The oldest and still best-written innovation value came from 3M’s William McKnight in 1948, as a management credo about delegation. The load-bearing sentence is worth quoting exactly: “Management that is destructively critical when mistakes are made kills initiative. And it’s essential that we have many people with initiative if we are to continue to grow.” (The popular summary — “hire good people and leave them alone” — is a paraphrase; the original is sharper, because it names the mechanism of the kill: destructive criticism of mistakes, not restriction of freedom.)

Notice what the sentence is: a tolerance clause — a written, senior-signed prediction of what will happen to you when your initiative fails. That’s precisely the private forecast every would-be intrapreneur is running, answered in advance by the founder. The culture that clause protected produced the 15% time that produced the Post-it, decades later — the through-line I traced in the intrapreneurship post. A values statement that wants to permit innovation needs its own tolerance clause, and it needs the same properties as McKnight’s: it predicts management’s behavior, not employees’; and it’s falsifiable — one public punishment of an honest failed bet breaks it, visibly, which is what makes keeping it meaningful.

The necessary edit for the modern version comes from Amy Edmondson’s failure taxonomy, which I use as a grading rubric in the innovation lab post: the tolerance clause covers intelligent failures — new territory, hypothesis-driven, sized to the learning — and explicitly doesn’t cover sloppiness in known territory. A blanket “we celebrate failure” value is as corrosive as none, because everyone can see it protecting negligence, and a value everyone can see being gamed converts to cynicism at the full exchange rate.

Values as operating documents: Netflix and Amazon

The two most-copied culture artifacts of the last two decades are both, at bottom, values sections that escaped the lobby poster and became operating documents — and both are studied by innovation people because they encode permission to move without asking.

The Netflix culture deck — 127 slides posted publicly in 2009, written by Reed Hastings and Patty McCord, reportedly called the most important document ever to come out of Silicon Valley by Sheryl Sandberg, and viewed well over seventeen million times — is remembered for its severance line (“adequate performance gets a generous severance package,” verbatim from the keeper-test section). But its innovation payload is the freedom-and-responsibility loop: increase talent density until you can decrease process, because process is scar tissue that optimizes for preventing the last error, and error-prevention culture is precisely what a business needs until it needs to invent something. The deck is a values statement with a theory of why these values, which is what makes it teachable rather than decorative.

Amazon’s Leadership Principles do the same work in a different register — sixteen principles maintained on a public page and, crucially, wired into the interview loop, the promotion process, and the vocabulary of everyday debate. Three of them are a compact innovation permission structure: Invent and Simplify (which includes the remarkable phrase “as we do new things, we accept that we may be misunderstood for long periods of time” — the patience value from the ten-year target post written into HR machinery), Bias for Action (“many decisions and actions are reversible and do not need extensive study” — a values-level endorsement of cheap experiments), and Are Right, A Lot (which Amazon glosses as seeking diverse perspectives and working to disconfirm your beliefs — a values-level endorsement of the kill decision). Add Bezos’s Day 1 doctrine — the 2016 shareholder letter’s “Day 2 is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death.” — and you have a company whose stated values would let you predict its experiment volume. That’s the test: could a stranger, reading only your values, predict how your company treats a failed bet? At Enron the prediction failed catastrophically. At Amazon it mostly doesn’t.

Writing yours: three clauses innovation needs

Sit down with Wickman’s three-to-seven constraint and the innovation literature, and the values an innovation strategy actually depends on compress to three clauses — write them in whatever house language fits, but make sure each survives the enactment test.

A tolerance clause (McKnight’s): what happens to intelligent failure here, stated as a promise about management’s behavior. Enacted test: name the last person whose failed bet was followed by a promotion. If the answer is nobody, the clause is espoused only — and every employee already knows it.

A candor clause (Grove’s and Edmondson’s territory, which gets its full treatment in the issues-list post): bad news travels fast here, and the messenger is safe. Enacted test: how did the last person who killed their own project — or contradicted a senior sponsor with data — fare?

A patience clause (Amazon’s “misunderstood for long periods”): we hold declared long bets through their ugly middle years, against stated kill conditions rather than mood. Enacted test: point to a current bet that looked bad last year and is still funded because its milestones say it should be — not because nobody dared raise it.

Then use them the way EOS insists and almost nobody does: in hiring (“tell me about a bet you lost”), in reviews, in the public narrative around every kill and every graduation. Values enter the V/TO first not because they’re inspirational but because they’re the cheapest layer of the whole stack — three sentences, no budget — and the only layer that operates in the thousands of moments when no process is watching. The rest of the organizer decides what you’ll bet on; this section decides whether anyone will dare to bring you the bet at all. Next up the V/TO: core focus — deciding what business your bets are even allowed to be in.