Eleven years into the same study, the same gap. The 2025 edition of Women in the Workplace by McKinsey and LeanIn, drawing on data from over 120 companies and 9,000 employees, restates the finding that the rest of the pipeline conversation tends to skip over: the weakest joint in the corporate ladder is not the boardroom, it is the very first management promotion.
The numbers, in their own words: "For every 100 men promoted to manager last year, only 93 women were promoted, and the numbers are worse for women of colour: just 82 Asian women and Latinas and 60 Black women were promoted." This is what the report calls the broken rung, and it is the structural reason every level above it stays disproportionately male.
Why The Number Sits There
Promotion to first-line manager is the first time a behavioural judgement, not just a performance one, gets formalised. Up to that point, an individual contributor's record is mostly about output. At the manager gate, the conversation shifts: can this person hold a room, give difficult feedback, take a stretch decision under uncertainty, push back on senior peers when needed? Those are behavioural assessments, and they are far more sensitive to bias, to who got which stretch assignment, to who was sponsored versus who was mentored, to who was already in the corridor when the role came up.
The 2025 report adds a second finding that should change how organisations read it. For the first time, McKinsey and LeanIn report a notable ambition gap: 80% of women want to be promoted to the next level, compared with 86% of men. The temptation is to treat that as a confidence problem, or worse, as a personal-choice story. The data does not support that reading. The same report notes the gap closes when women receive the same career support as men: sponsorship, manager support, and access to stretch opportunities.
In other words, the ambition gap is downstream of the experience gap. It is what people want when the system has already taught them what to expect.
What Most Organisations Do
The default response to a broken-rung number is broadly the same one organisations have used since the first Women in the Workplace report in 2015: a sponsorship programme, an unconscious bias module, a leadership cohort for high-potential women. These are not bad ideas, but the evidence on each is uneven, and the most common interventions are precisely the ones least likely to change a single behaviour at the moment of promotion.
Unconscious bias training, in particular, has a thin evidence base when measured against observed behaviour change. Awareness rises, sentiment improves, and the manager who would have promoted the louder candidate still promotes the louder candidate three months later. Roediger and Karpicke (2006), in Psychological Science, found that active retrieval improves long-term retention by around 50% compared with re-reading. Awareness sessions are re-reads. Promotion calibration in front of a sceptical observer, with feedback, is retrieval.
What Works At The Promotion Gate
The change is small in scope, large in leverage. The behaviour to target is not a 90-minute online module. It is the moment a manager makes the case for one candidate over another in a calibration meeting, and the moment a sponsor decides whether to put their political capital behind someone whose progress is invisible to them.
Real behaviour change happens through lived experiences that make memories stick. Our work with leadership groups uses immersive simulation to surface those exact moments, calibration meetings staged with professional actors playing the affected stakeholders, complete with the political cross-currents that make real meetings hard. Leaders practise the conversation they will actually have to have, twice, with feedback. In our own academic research, building on behaviour-change work from UCL, Cambridge and Bocconi, immersive role-play produced approximately 20% higher observed behaviour change than passive modalities such as slides and video. Self-rated confidence, importantly, did not predict observed performance.
The implication for an organisation looking at the McKinsey numbers is that the broken rung is fixable, but not by repackaging the awareness layer. It is fixed at the gate, in the rooms where promotions are decided, by managers who have rehearsed the moves they need to make.
Inclusive leadership is not an attitude. It is a set of small, observable behaviours, repeated under realistic conditions, until they become the default in the meeting that matters. The McKinsey number is not stuck because the topic is hard. It is stuck because the rehearsal has not happened.
Book a free 30-minute diagnostic call → or read about our work with companies.