The resignation lands in month twenty-two and everybody is surprised by it. She was well regarded. There had been no complaint. The exit interview records a personal decision and a family consideration, and the file closes.
The decision was not made in month twenty-two. It was made across about ninety days after returning from maternity leave, nearly two years earlier, and nobody in the organisation was watching while it happened.
The 2026 AIMA and KPMG in India Women Leadership Survey puts it precisely. Women’s exits are rarely driven by declining aspiration. They reflect accumulated organisational frictions, role design constraints, and uneven support during critical transitions.
In Indian organisations, the transition that goes unsupported is the return from maternity leave.
What actually happens in those ninety days
Most of it has been normalised over decades, which is exactly why it goes unmanaged.
She comes back. The account she ran has been split between two people, and both of them have now been doing it for six months. Everyone agrees she will pick it back up, and nobody has said when or how much. In the meantime she is given a set of tasks that are undeniably real and but peripheral.
The interesting project started in month four and is well underway. Nobody excluded her from it. She simply was not there when it was staffed, and restaffing a live project is disruptive, so the question never gets asked.
Her manager, who is caring and means well, is being careful not to overload her. That care expresses itself as fewer asks, lighter work, no travel. She experiences it as being set aside.
None of this is hostile, or even intentional. Every individual decision in that sequence is defensible. And by day ninety, a version of her role has settled that is smaller than the one she left, and everyone has adjusted to it, except in her mind.
The resignation arrives eighteen months later and gets recorded as a personal decision.
What this has to do with your leadership pipeline
More than it looks, and this is the part I would want a talent function to sit with.
Every women’s leadership programme operates on whoever is still there. It draws from the manager and senior manager population as it exists on the day nominations open, and it can only develop the people in that pool. Which means anything that removes women from the pool before they become eligible is, in practice, a selection mechanism sitting upstream of your entire development investment.
And that is what an unmanaged return is. It is not neutral attrition. It filters on whether a particular manager happened to handle a particular ninety days well, and has nothing to do with capability, ambition or potential. The women who come through it with their trajectory intact are the ones who had a good manager that quarter.
So you can commission an excellent programme, deliver it well, and still watch senior representation refuse to move. Not because the programme failed. Because the pool it drew from had already been shaped by something nobody was managing.
Return design is talent selection, retention and pipeline building. We have simply never called it that, and so it never got that attention.
Why this transition specifically
Because it is the only major career transition that most organisations treat as an absence rather than a change.
We onboard external hires. We increasingly onboard internal promotions. We have almost nothing for the person coming back, because the mental model is that she was here, then she was away, and now she is here again. Nothing to design.
But she is returning to an organisation that has moved, into a role that has been redistributed, with a life outside work that has structurally changed, and with her own confidence in a place it has not been for years.
The design that changes the outcome
What I have seen work is unglamorous and largely free.
A portfolio conversation in week one, with a named date for full restoration. Not a vague assurance that things will get back to normal. A specific agreement about what returns, when, and who is accountable for making sure it happened.
A staged ramp that is agreed rather than assumed. The difference between a plan she participated in and a plan applied to her is the difference between support and sidelining, and from the outside those two look identical.
One protected piece of visible work in the first quarter. Something that puts her in front of people who matter, chosen deliberately. This single item does more for trajectory than anything else on the list, because it interrupts the story the organisation is otherwise starting to tell about her.
And a check-in at day ninety with someone who is not her manager. Her manager is usually the person whose well-intentioned decisions created the drift, and is therefore the least able to see it.
The number that would settle this internally
If you want to know whether this is happening in your organisation, do not run a survey.
Plot your regretted attrition of women against months since return from maternity leave. Not against grade. Not against tenure. Months since return.
Most organisations have never cut the data that way, and in most of them the pattern is visible immediately. It is a query, not a project, and it converts a conversation about culture into a conversation about a specific window with a specific cost attached.
The ambition was never the issue.