When a woman leaves tech without leaving her employer, the retention statistics never see it.
Attrition figures measure resignation dates. They cannot measure the analyst who moves into a project-management role with no technical content, or the developer who accepts a people-management promotion specifically because the alternative was being managed out. Both women remain on headcount. Neither is developing a technical career.
An employer whose senior technical women are quietly absorbed into the PMO is not failing any published metric.
This kind of exit — sideways into a non-technical function, or upward into pure management — is structurally different from leaving an employer, but its effect on the technical pipeline is nearly identical. The woman is gone from the discipline. The gap she leaves in the senior technical cohort is real. The retention statistic does not record it.
The mechanism is well documented even if rarely counted. Mid-career women in technical roles report being steered — sometimes explicitly, sometimes through the slower pressure of being overlooked for technical leads in favour of junior men — toward roles that are coded as relational: coordination, liaison, programme governance. These roles are not necessarily less paid or less senior. They are, however, not technical. Once a woman has spent two or three years in one, her path back to deep technical work is substantially harder than it would have been had she stayed.
The framing matters because mid-career attrition tends to be analysed as a retention problem when it appears in resignation data, and largely ignored when it does not. An employer whose senior technical women are quietly absorbed into the PMO is not failing any published metric. Its gender pay gap may even narrow, because the women are moving into roles whose pay bands sit above entry-level technical grades. The gap number improves; the technical pipeline empties.
- Resignation data
- counts formal departures only; internal moves are invisible to it
- Gender pay gap reporting
- a distribution comparison across the whole employer; a sideways move into a better-paid non-technical role can improve it while hollowing the technical cohort
- Exit interview data
- only collected from people who leave the employer; internal movers are never asked
- Internal mobility data
- the one source that could track this; rarely published or analysed by gender and function together
There is also a compounding effect with part-time working and flexible arrangements. Women who return from parental leave and request reduced hours are disproportionately likely to find that the only roles offered at reduced hours are operational or managerial rather than technical. The accommodation is real, but it is applied in a way that removes the employee from her specialism. Within a few years the specialism has lapsed, and the route back requires reskilling that nobody is formally asked to fund.
None of this appears in exit interviews, because there is no exit. It does not show in the employer's published reporting, because the headcount denominator stays constant. It is auditable, in principle, from internal mobility data — which functions women move into, at what career stage, and whether those moves are reversible — but almost no published analysis tracks it. The result is a structural loss that the instruments designed to measure retention are not designed to see.