The Glass Ceiling Is a Set of Decisions
When capable people from underrepresented groups stall, leaders tend to blame "the system." That word hides four specific choices. Each one can be made differently.
The Problem Organisations recruit women, LGBTQ+ professionals and persons with disabilities into the pipeline, then watch them stall below senior levels. Leaders read this as a supply problem or an individual one.
Why It Happens Four structural failures hold the ceiling in place: the real rules of advancement are never shared, the work of advancing is assigned entirely to the individual, nothing checks the judgment of evaluators, and "the system" is used as an alibi that detaches outcomes from the people who produced them.
The Solution Treat the promotion structure as a set of decisions with owners. Publish the operative rules, supply sponsorship and visibility as part of the structure, build a layer that checks judgment, and assign a named owner to every decision that shapes advancement.
Most leaders I work with do not believe their organisations have a glass ceiling. They point to the policy, which is neutral. They point to the promotion criteria, which apply to everyone. They point to hiring numbers, which have improved. Then they look at the leadership team and cannot explain why it looks the way it did ten years ago.
The explanation is rarely in the policy. It is in how decisions actually get made, and in the fact that almost nobody is asked to account for them.
From the individual's side, the experience is consistent. You get into the room, which was hard enough. You hit the targets. You do the job you were given and the job nobody bothered to write down. You are told the sky is the limit, so you look up and climb, and you hit something solid. You climb again and hit it again, in the same place. Because you cannot see it, you cannot name it, and you start to wonder whether the fault is yours.
The standard advice agrees that it is. Be more visible. Be more patient. Network harder. Find a mentor. That advice is comfortable for everyone except the person receiving it, because it keeps the problem on their side of the table.
The glass did not get there by accident. In my experience leading inclusion work in Indian financial services, it is held up by four failures, and people meet them in roughly this order.
Failure 1: The Operative Rules Are Never Shared
Every organisation runs on two sets of rules.
The first set is written down. It lives in the policy, the competency framework and the published path to the next level. Because it is given to everyone, it looks like the whole picture.
The second set is the operative one. It governs how advancement actually happens: which projects carry weight, which relationships matter, how visible you need to be and to whom, what a promotion committee quietly reads as "ready." This set is almost never written down, because writing it down would invite questions about it.
Professionals who match the organisation's default absorb the operative rules through proximity. Their managers look familiar. Their networks are already open. The environment was calibrated to people like them before they arrived. They learn the real game without noticing they are learning it.
Everyone else receives the written rules and is left to reconstruct the operative ones alone, usually late, usually after a cost has already been paid. A woman returning from maternity leave discovers the high visibility account was reassigned and that visibility, not the account's revenue, was what the committee valued. A gay manager who keeps his personal life out of conversations finds he has also kept himself out of the informal circle where sponsors are made.
Opacity is never neutral. It rewards whoever was close enough to absorb what was never said.
What leaders can do: Interview the last ten people promoted into senior roles and ask what actually got them there. Compare their answers with the written criteria. Where the two differ, either publish the operative rule or retire it. If a rule cannot survive being written down, it should not be deciding careers.
Failure 2: The Whole Weight Sits on the Individual
A structure that will not explain itself also will not carry itself. It turns progression into a personal project and hands the project to the professional.
Consider what the organisation expects an underrepresented professional to source and sustain alone:
Resilience, enough to absorb setbacks the structure created and treat them as personal tests
Visibility, manufactured deliberately, in rooms that do not extend it by default
Sponsorship, a senior advocate they must find and earn inside networks not built around them
Perception management, the constant labour of monitoring and correcting how others read them
Research has made the sponsorship gap in particular hard to ignore. Herminia Ibarra and her colleagues found that women were often overmentored and undersponsored: they received advice, while men more often received advocacy that translated into roles. Sylvia Ann Hewlett's work on sponsorship reached a similar conclusion. Mentors tell you how to navigate. Sponsors spend their own credibility to put you in the room.
When sponsorship is left to chance, it follows familiarity. And familiarity follows the default.
A system that places its entire load on the individual has made a decision. It has decided to carry none of the load itself and to call the outcome merit.
What leaders can do: Make sponsorship a formal responsibility of senior leaders, with named sponsees and outcomes that are tracked and reviewed. Measure who gets stretch assignments, client exposure and presentation time with the board, broken down by group. If visibility is a criterion for promotion, the organisation is responsible for distributing it.
Failure 3: Nothing Checks the Judgment
This is the failure that matters most, and the one almost no organisation has addressed.
Underrepresented professionals often differ from the default an evaluating room is calibrated to. They may speak differently, carry a different background, or hold a different idea of what confidence or readiness looks like. A person with a disability may have a working pattern the room has never seen at the senior level. None of this has anything to do with capability. All of it affects how capability is read.
Bias enters every judgment made by one human about another. The useful question is not whether it enters, but what exists to catch it before it hardens into a verdict.
In most organisations, nothing does. One evaluator forms one read, and that read becomes the outcome. There is no second opinion with standing to disagree, no record of why the decision was made, and no later check on whether decisions cluster in a pattern.
Iris Bohnet's work on behavioural design makes the case that debiasing individuals is far less reliable than redesigning the process around them. A structure built for fairness includes four checks:
Criteria fixed before evaluation, written and applied to everyone, so the standard cannot move once a particular candidate meets it
More than one evaluator, with genuine authority to disagree, so a single read is never the whole verdict
A documented rationale for every decision, so judgment can be examined rather than assumed
A pattern audit over time, because one biased decision is hard to see, and the same outcome landing on the same group again and again is not
Without that layer, a difference that should have been irrelevant becomes the thing that decided. Nobody can prove it, which is exactly why it persists.
What leaders can do: Introduce all four checks into the next promotion cycle, starting with the documented rationale, which costs almost nothing. Then run the audit on the last three cycles. The patterns are usually visible within an afternoon.
Failure 4: "The System" Becomes the Alibi
Ask why a capable underrepresented professional did not advance, and the answer arrives in a familiar shape. The system. The process. How things work here. The way the cycle runs.
A system is a set of decisions made by identifiable people. Someone chose not to publish the operative rules. Someone chose to assign every instrument of advancement to the individual. Someone chose to let a single unchecked read stand as a verdict. Each of those is a choice with a person attached.
Calling it "the system" detaches the choices from the people who made them. It turns decisions into weather, something no individual can be expected to answer for. The professional is told, in effect, that nothing went wrong. The system simply ran.
This is why the fourth failure has to be named before the other three can be fixed. As long as the structure is described as an object, nobody is accountable for it. The moment it is described as a set of decisions, everyone who made one is.
What leaders can do: List every decision point that shapes advancement, from project allocation to calibration meetings to succession slates. Assign a named owner to each. When outcomes are reviewed, the question shifts from "what did the system produce" to "who decided, and on what basis."
Where to Start
Leaders who want to act on this do not need a new programme. They need to change four decisions they already make.
Publish the operative rules, so advancement is a known path and stops being an inherited secret.
Supply the instruments of progression, so sponsorship and visibility are part of the structure.
Build the layer that checks judgment, so no single read becomes a verdict and patterns get examined.
Put names against decisions, so "the system" can no longer answer on anyone's behalf.
None of this is new. Organisations have written their rules down, formalised sponsorship and structured their evaluations, and seen results. The method has never been the obstacle.
The obstacle is whether the people with the authority to change these decisions want to. That is the question every leader reading this has to answer, and it is a simpler one than it looks.
Do you care?
Maira Q. writes on inclusion, identity, and belonging. Read more in Brewing Inclusivity on LinkedIn or subscribe to Letters from Maira on Substack.
Comments