I spent the first 20 years of my career in investment banking – an industry where numbers aren’t just important, they’re everything. You live and die by the accuracy of your models. You’re trained to interrogate data, to question patterns, and to distinguish genuine volatility from systemic failure.

In any organisation, if a CEO wanted to reassure shareholders, they turned to the numbers. And if a CFO presented results that were inconsistent, illogical, or repeatedly trending in the wrong direction, alarms would go off immediately. No board would accept “random variation” as an explanation; they would demand clarity, action, and accountability.

That’s precisely why the latest 2025 Women in the Workplace data is so striking. When you apply the same analytical scrutiny we apply to financial performance, the corporate pipeline numbers tell a story that cannot be dismissed as chance.

A Pattern Too Consistent To Be Random

At entry level, women make up around 49% of employees. If progression were fair and unbiased, you would expect senior levels to reflect that – perhaps with small fluctuations, but broadly stable.

Yet the representation of women declines at every promotion point:

• Manager: 42%
• Senior Manager/Director: 39%
• Vice President: 35%
• Senior Vice President: 32%
• C-suite: 29%

Meanwhile, white men rise from 32% at entry level to 56% of the C-suite.

This pattern is not jagged or volatile – it is steady, predictable and reinforced year after year across thousands of companies.

If a CFO presented results that consistently drifted in one direction without any natural variation, they would not be allowed to shrug and say “that’s just how the numbers landed.”

A Thought Experiment: If Promotions Were Random

To be clear: promotions are not random. No one genuinely believes they are.

However, probability is a useful thought experiment because it reveals how unlikely this pattern would be if the system were neutral. If promotions operated without bias, you would expect:

• some years where women moved up at higher rates
• some years where men did
• and a long-run average close to the starting composition

Instead, we see a one-directional trend.

For randomness to produce this, the system would need to:

• consistently favour men at every single level
• across every industry
• across every year
• while systematically reducing the representation of women, especially women of colour

The probability of this happening by chance is extraordinarily low – far lower than most people imagine.

It is the statistical equivalent of flipping a coin repeatedly across multiple companies and getting tails almost every time.

If these were financial results, not talent outcomes, they would immediately trigger a board-level investigation.

Addressing The Common Objections

A balanced analysis must consider alternative explanations. So let’s address them directly.

Objection 1: “Different industries attract different genders.”

True – but irrelevant to the core argument.
Even when you examine specific industries, functions or job families, the downward pattern remains remarkably consistent.

Objection 2: “Women opt out at higher rates.”

The data shows otherwise. Women remain ambitious at the same rate as men.
What differs is access: sponsorship, visibility, stretch roles, and how performance is assessed.

Objection 3: “This isn’t intentional discrimination.”

Exactly.
Most of this is not intentional – but systems don’t need intent to produce unequal outcomes.
Legacy structures, hidden norms and biased evaluation criteria compound silently over time.

Objection 4: “Maybe the numbers fluctuate year to year.”

They don’t.
The trend is stable across a decade of data. If this were a financial metric, consistency of decline would itself be the smoking gun.

Acknowledging these objections doesn’t weaken the argument – it strengthens it. Because once you examine them carefully, none can fully explain the size, consistency or direction of the trend.

So What Do The Numbers Actually Tell Us?

Once you rule out randomness, individual preference, or isolated industry effects, the explanation becomes clear:

The corporate pipeline is functioning exactly as its systems, structures and incentives shape it to function.

The 2025 data reflects:

• uneven access to stretch opportunities
• inconsistent sponsorship
• evaluation systems that reward some behaviours and penalise others
• cultural norms that advantage specific groups
• systemic barriers that accumulate over time, especially for women of colour

These are not accidents; they are outcomes.

If Numbers Matter, They Must Matter Everywhere

Businesses demand accountability in financial performance. They operate on dashboards, metrics, risk models and forecasts. If numbers matter there, they must also matter in people systems.

You cannot celebrate numerical discipline in the P&L while tolerating statistical impossibility in talent progression.

Because here’s the truth:

If a CFO couldn’t keep their job with trends this inconsistent and this irrational, the system that produces our leadership demographics shouldn’t keep its job either.

Talent is the only asset that appreciates in value – but only when the structure around it is designed to be fair, consistent and intentional.

The 2025 Women in the Workplace data is not a warning sign.  It’s a diagnosis.

And like any good analyst, once you understand what the numbers are telling you, you can no longer claim it’s random.

If this resonates with you, I’d ask you to share it or comment. The more we bring a numbers-driven lens to this issue, the harder it becomes to ignore – and perhaps a purely numerical approach is exactly what’s needed to get more people thinking seriously about it.