What the research found
Researchers at the University of Milano-Bicocca and Radboud University tested 120 university students on two distinct types of problems. One was a classic “insight” puzzle, solvable only through a sudden mental leap. The other was a step-by-step problem with a clear, logical solution path. Half the participants were asked to think aloud continuously while working; the other half solved their problem in silence.
The results were near-mirror images of each other. Among those solving the insight problem, only 13% succeeded while narrating their thinking out loud, compared to 57% who solved it in silence. For the step-by-step problem, the pattern flipped entirely: 80% succeeded while talking it through, versus just 30% working silently.
The researchers’ explanation: the two problem types rely on different mental processes.
Step-by-step problems are already conscious and orderly, so verbalizing helps organize them.
Insight problems depend on an unconscious restructuring — and forcing that process into words, in real time, consumes the very cognitive resources restructuring requires, effectively jamming the signal before a solution can surface.
The Amazon parallel
Years before this study existed, Jeff Bezos had already built Amazon’s meeting culture around a strikingly similar instinct. As Fortune documented in a 2012 profile naming Bezos its Businessperson of the Year, meetings of Amazon’s senior leadership team begin not with discussion, but with everyone silently reading a written memo together — a practice so unusual that new executives are often visibly thrown by it.
Bezos has said the shared silence is deliberate: it guarantees the group’s full, undivided attention before anyone speaks. He’s also pointed to writing itself as the real discipline — arguing it’s essentially impossible to produce a clear, well-structured six-page memo without having fully thought the problem through.
In later interviews, Bezos has described his objection to traditional slide presentations in terms that echo the research almost exactly: a senior executive interrupts a presenter on slide three, before the idea has had room to fully form.
Amazon’s fix was structural — separate the silent, individual thinking phase from the verbal, collective one, rather than asking people to do both at once.
A second case: withholding conclusions
A related instinct shows up in Pixar and Disney Animation president Ed Catmull’s account of taking over Disney Animation’s leadership, in an interview with McKinsey. Rather than moving quickly to diagnose problems — the instinct most incoming executives are praised for — Catmull describes spending two full months simply observing and coaching individually, deliberately drawing no conclusions about people or process.
The goal, he says, was to let the underlying psychology and sociology of the organization become clear before forming — let alone announcing — any judgment.
It’s the same discipline applied to leadership rather than individual cognition: resisting the pressure to verbalize a conclusion before the underlying pattern-recognition process has finished.
Catmull is explicit about the cost of skipping that step — he points to Disney Animation’s own creative decline after Walt Disney’s death as a case of leaders drawing fast, confident conclusions from early success, then locking the studio into a formula that eventually failed.
.
.
Why it matters
The finding cuts against a common workplace assumption: that transparency of thought — narrating decisions, thinking aloud in meetings, presenting live — always improves outcomes.
The research, together with two decades of Amazon’s operating history and Catmull’s own account of organizational turnarounds, suggests that assumption holds only for problems with a clear, linear path to a solution.
For the genuinely novel problems that demand a creative leap — a stalled strategy, an unread culture, a market position no one has quite figured out — real-time narration may get in the way, and quiet, undistracted thinking time may serve teams better than another round of talking it through.
The study’s authors note their findings come from controlled laboratory puzzles rather than workplace scenarios, and call for further research using real-world tasks. But the fact that two of the most studied operators in modern business independently arrived at a comparable practice — one through a memo-reading ritual sustained for over twenty years, the other through a deliberate two-month silence at a moment when speed was expected — suggests the lab and the boardroom may be describing the same phenomenon from two very different vantage points.
