A leader who never took an accounting class can still read a P&L cold. Revenue up, margin down, cost of goods creeping. The eye goes straight to the number that moved and asks why within seconds. That fluency was built on purpose, over years of dashboards and quarterly reviews, in a system where the people building those instruments were, for most of corporate history, almost entirely men. As of 2026, women hold roughly eleven percent of Fortune 500 CEO seats. Go one level down, into the P&L-owning roles that are the actual credentialing path to that seat, and the picture sharpens further: women hold about a quarter of CFO and COO positions, and under nine percent of CTO roles. The instruments a company runs on were built almost entirely by people in one demographic, tracking what mattered to them.

‍

This is what the book argues from its opening page: the friction women hit at work is not a personal failing; it is a system that was never built to register certain results as results. Nowhere is that more literal than in what a leader’s eye has actually been trained to see. And seeing has to come first. You cannot name a pattern you have never registered as a pattern, and you cannot redesign a system you cannot see. Everything that follows in this piece is about that first, hardest step, before any redesign, whether that means who gets to decide, who gets held accountable, how resources move, or what gets treated as a signal worth acting on, is even possible.

‍

Put one of those leaders in a meeting where a woman is cut off mid-sentence by a more junior man. It happens once, then twice, then a third time before the meeting ends, and each time the room simply moves on, because nothing about an interruption looks like the kind of event that leader has been trained to log. A margin drop trips an alarm built over years of dashboards, forecasts, and quarterly reviews designed for exactly that purpose. An interruption trips nothing, because no one in that leader’s career ever built the equivalent alarm for it. Three data points that would end a meeting early if they showed up on a P&L pass by unremarked, folded into what the leader will later describe, honestly, as a normal Tuesday.

‍

That is not a coincidence of training. It is a consequence of who was in the room when the training got decided. Kieran Snyder, an empirical linguist who later founded Textio, first measured the interruption pattern itself in 2014 by sitting in on real workplace meetings. She found men interrupted at roughly twice the rate women did and were about three times more likely to interrupt a woman than to interrupt another man. A decade later, after logging more than a thousand additional hours of meeting data, Snyder found seniority was the sharper predictor. More senior people interrupted more junior people, consistently. In a leadership layer that is still overwhelmingly male, seniority and gender are not two separate variables. They compound.

‍

Shelley Correll, at Stanford, found the deeper mechanism behind why this never gets logged. Her 2020 study in the American Sociological Review, built around what she and her coauthors call the Viewing and Valuing model, found that managers do not evaluate everything an employee does. They evaluate what they have been trained to notice. Training is not neutral. It is built by whoever held the pen when the evaluation criteria were written, and for most of the history of corporate performance management, that pen was in the hand of someone who had never personally been talked over in a meeting and watched his idea get credited to someone else thirty seconds later. He had no lived reason to think that moment was worth instrumenting, so he did not build the instrument. Not out of malice. Out of the simple, ordinary limit of designing systems around your own experience.

‍

Here is the part that turns this from a diagnosis into a lever. Christopher Karpowitz at BYU and Tali Mendelberg at Princeton analyzed eighty-seven real school board meetings across twenty states and found that women spoke as often as their male colleagues only on boards where women held at least sixty percent of the seats. Below that threshold, women used a fraction of their proportional share of speaking time. Their companion lab research, published in the American Political Science Review, found the identical shape: under majority-rule decision making, women in the numerical minority absorbed a net negative balance of interruptions. Once women made up the clear majority of the room, that balance reversed. This is not a personality fix, and it is not solved by asking anyone to be nicer. It is solved by decision architecture, specifically, by who is in the room and what rule governs how the room decides. That is not a training problem. It is a composition problem, and composition is something a leader with hiring and promotion authority can actually change.

‍

That same composition problem is deciding who gets promoted in the first place, long before anyone reaches the top, which is exactly why the top still looks the way it does. LeanIn.org and McKinsey named this the broken rung in their 2019 Women in the Workplace report, now in its second decade of data: the biggest gap in the entire leadership pipeline is not at the C-suite. It is at the very first promotion, from individual contributor to manager. In the most recent report, for every hundred men promoted into that first management role, only ninety-three women made the same jump, with the gap widening further for women of color. That decision gets made in thousands of small, unremarkable rooms, at exactly the composition ratios where the interruption pattern is most severe.

‍

Here is what that costs, in the currency every one of these leaders already tracks. SHRM and Gallup both put the price of replacing an employee at fifty to two hundred percent of that person’s annual salary, once recruiting, ramp time, and lost institutional knowledge are counted. A capable woman who gets talked over in every meeting for two years, watches a colleague get promoted into the role she was doing the work of, and leaves, does not appear on any dashboard as a pattern. She appears as a single resignation line, absorbed in an unremarkable turnover number, budgeted for and shrugged off. The leader who cannot name why she left will authorize a six-figure replacement cost without ever connecting it to the meeting eighteen months earlier where she was interrupted three times and no one said a word.

‍

This is not a story about leaders who could not have known. The interruption happened in front of them. The promotion decision was theirs to make. The composition of the room was theirs to change, and the research on exactly how to change it already exists. What was missing was never the information. It was the instrument, and the instrument was never built because the people with the authority to build it had no personal stake in needing it. That is a design flaw with an origin, not a mystery, and once it has an origin, it has an owner, and now, a documented fix.

‍

That is why seeing is the hardest part of any rebuild. Not because the problem is hidden, but because the people holding the pen were never the ones it happened to, so it never occurred to them to write it down. Once a leader understands that the blindness has a traceable cause and a measurable threshold where it reverses, “I did not notice” stops being an explanation and starts being the design flaw itself, sitting on the books as a line item signed off on without knowing what it was actually for. You cannot unsee that once you have traced it, and the responsibility was never abstract. The organization head, the CEO, whoever holds the authority to redesign decision rights, evaluation criteria, and informal power structures, is the one who owns what happens next.