Launched early and underresourced, a decentralizing initiative is left to fail, and the executive behind it cites the failure as proof the organization was not ready and takes back control.
An example
Under pressure from its private-equity owner to move faster, a commercial cleaning company announced that its forty branch managers would control their own pricing and hiring, a change the company's president had opposed for years. He gave the kickoff speech. The branch managers got pricing authority and a three-week training course, while their hiring approvals still went through a regional office that took a month to respond. The pricing tool arrived late and without the cost data branches needed to quote. Within two quarters, margins fell at a third of the branches. At the board meeting the president presented the results as evidence that branch managers "aren't yet ready to run P&Ls," and the board approved a central pricing desk under his control. Nobody put the resource record next to the results.
Why it works
One visible, concrete failure discredits the whole case for decentralizing, its speed, local judgment, and accountability, so the alternative looks worse. The post-mortem then credits central control with the benefits of discipline, coordination, and supervision. Anyone who wants to try again must explain why their version will not repeat the failure, and every later stumble is read against the precedent. The underfunding stays invisible only if the executive controls staffing and resources, the culture reads outcomes without examining the conditions of the attempt, and the executive appears supportive throughout. The failure also has to be visible enough to discuss and small enough to spare the sponsor any accountability.
The thinking behind it
Lee Ross Ross's The Intuitive Psychologist and His Shortcomings: Distortions in the Attribution Process (1977) names the fundamental attribution error: observers attribute outcomes to people's traits and underestimate their situation. Branch managers who fail without cost data, hiring authority, or a working tool are judged unready, and the starved conditions go unexamined. Putting the resource record beside the results, before anyone draws a conclusion, is the correction.
Jonathan Baron Baron and Hershey's Outcome Bias in Decision Evaluation (1988) found that people treat outcomes as evidence about decision quality even where they carry none. A failed decentralization is taken as proof that decentralizing was the wrong decision, whatever the conditions of the attempt. Pre-registered success criteria and resource floors, set before launch, give evaluators something to judge besides the outcome.
Warning signs
Public backing from the sponsor, a kickoff speech or a token budget transfer, never turns into delegated authority, headcount, or a protected timeline. Approvals still run through the old center, and tools and data arrive late or incomplete. The executive launching the effort has a record of preferring the opposite. When results disappoint, the post-mortem concludes the organization was not ready and proposes more central control.
How to defend yourself
Settle the conditions of the attempt before launch, because that is where the fight is.
- Insist on pre-registered success criteria and resource floors written into the launch: authority, headcount, budget, tools, and the date each arrives.
- Ask for independent monitoring of whether those floors are met, reported to the same body that will judge the outcome.
- Log, with dates, every resource that arrived late or never arrived.
- In the post-mortem, put the resource record beside the outcome before anyone concludes the organization was not ready.
If you run one of the units, find the parts you can make work with what you have. A team that delivers despite the starvation is the hardest evidence against the conclusion the launch was built to produce.
How a leader shuts it down
Separate who launches a structural experiment from who judges it, especially when the person launching it prefers the opposite outcome. Write resource floors and success criteria into the launch documents, and audit them before reading results. Make every post-mortem answer whether the attempt received what it was promised.