Here’s an uncomfortable truth that most leadership consultants won’t tell you: culture change programs almost never work.
The reason isn’t that culture doesn’t matter. It matters enormously. The reason is that most organizations treat culture as an awareness campaign when it’s actually a systems engineering problem. They hold offsites, print new values on the wall, launch internal communications campaigns about “embracing innovation” — and then send everyone back to the same approval hierarchies, the same incentive structures, the same meeting rituals that produced the old culture in the first place.
The organizations that actually break refusal culture don’t fix the culture. They redesign the system.
The most radical example is Haier, the Chinese appliance maker. In 2005, CEO Zhang Ruimin looked at his company’s middle management layer and saw the immune system we talked about in a previous article — thousands of people whose primary organizational function was filtering, approving, and rejecting. So he eliminated roughly twelve thousand middle-management positions and reorganized over eighty thousand employees into approximately four thousand self-managing microenterprises.
Each microenterprise has three rights: decision-making, hiring and firing, and profit distribution. Zhang’s explanation to McKinsey was characteristically blunt: “We removed the intermediate layer. After bureaucracy was eliminated, the organization was no longer hierarchical. It became a decentralized, distributed network.”
Haier’s revenue has grown to roughly fifty-seven billion dollars. The company reports twenty-three to twenty-eight percent annual profit growth over the Rendanheyi decade, compared to roughly one percent for traditionally structured peers.
You don’t have to be as radical as Haier. Adobe’s intervention was elegant precisely because it was small.
In 2012, VP of Creativity Mark Randall designed a program called Kickbox. The concept was deliberately simple: any Adobe employee could request a red cardboard box containing a one-thousand-dollar prepaid credit card, a notebook, a pen, a Starbucks gift card, and step-by-step instructions for testing an idea. No expense reports. No prior management approval. No pitch deck required.
Randall’s design principle cuts to the heart of refusal culture: “Every executive has the power to say yes, none of them have the power to say no.” He didn’t try to change how executives think about innovation. He simply removed their ability to kill ideas before those ideas had a chance to generate evidence.
In its first year, roughly one thousand Adobe employees went through Kickbox and tested nearly a thousand ideas — compared to twelve to twenty-four product prototypes per year under the old system. Twenty-three ideas earned next-stage funding. Adobe later open-sourced the entire program, and it’s been adopted by companies including Cisco and GE.
Netflix took yet another approach. Reed Hastings and Erin Meyer, in No Rules Rules, describe a philosophy they call “context, not control.” Instead of approval chains, Netflix gives employees full strategic context — financials, customer data, competitive landscape — and trusts them to make decisions. No vacation policy. No expense policy beyond “act in Netflix’s best interest.” The structural distance between idea and action is essentially zero.
The common thread across Haier, Adobe, and Netflix isn’t a shared philosophy about innovation. It’s a shared diagnosis of the problem: the structural distance between the person with the idea and the person with the authority to greenlight it is the single strongest predictor of how much refusal culture an organization carries.
Gary Hamel and Michele Zanini’s research confirms this. Their seven-thousand-respondent survey found that an unbudgeted sign-off takes an average of roughly twenty days in large firms. Twenty days. That’s twenty days in which enthusiasm dies, context shifts, competitors move, and the person with the idea learns — again — that the system isn’t built for them.
This is why culture speeches don’t work. You can’t talk your way past a twenty-day approval process. You have to redesign it.
Pick one bottleneck in your organization’s innovation process — one approval, one committee, one sign-off — and run a thirty-day experiment without it. Track what happens. If the sky doesn’t fall, you’ve just found a piece of your immune system that was protecting nothing.