How to measure it, analyze it, and turn it into decisions that impact results.
“We have 48% women in our workforce.”
Good.
Now what?
Because diversity is not a snapshot, nor is it only about gender. It’s a broader variable that we can measure—and also design.
But not all diversity has the same impact.
The diversity that impacts performance is cognitive diversity: different ways of thinking, interpreting, and solving problems.
Other dimensions—such as identity—matter.
But they don’t always explain outcomes.
How do we observe cognitive diversity?
Cognitive diversity is not directly observable.
It is not defined by demographic variables.
It is approximated.
Through how people:
-
interpret problems
-
prioritize decisions
-
execute
-
analyze
-
project
A practical way to approach it is through competency profiles.
Not to measure whether someone is better or worse, but to understand how they stand out differently compared to others.
Cognitive diversity is not something you have.
It’s something you build.
What science says (and what it doesn’t)
The work of Scott E. Page (2007), The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies shows that:
Under certain conditions, groups made up of people with different perspectives, heuristics, and interpretations can outperform groups composed only of “brilliant” but homogeneous individuals.
Can under certain conditions. Not always. Not automatically.
The advantage is not in diversity as a slogan, but in the fact that different ways of thinking allow problems to be explored from multiple angles and help avoid collective blind spots.
The combination of sufficient competence + diversity of approaches can generate better results than the sum of excellent but similar profiles.
And those different approaches do not come from demographic variables.
They come from how people think… and how you choose to combine them.
The hard variable is not diversity. It’s how you manage it.
Without structure, diversity creates friction.
With structure, it creates innovation.
The elegant mistake
Confusing representation with impact.
Representation is counting.
Impact is measuring differences in outcomes.
If we don’t analyze how performance, engagement, development, or promotion are distributed across different profiles, we are not managing diversity—we are just reporting headcount.
What happens when you truly analyze it
The core idea is simple: whenever you measure something in your company—performance, climate, potential, onboarding—you can add the “diversity” variable to your analysis and let the data speak.
How to incorporate diversity into analysis?
-
Comparing performance percentiles by gender, age, or tenure (not just averages)
Even if we analyze results by gender, age, or tenure to detect patterns, cognitive diversity is individual. It is not determined by demographic variables. These help detect patterns, not explain how performance-driving diversity is created.
-
Analyzing strategic indicators by population (e.g., leadership, innovation, results orientation)
-
Segmenting with AND/OR clauses to detect invisible patterns
(e.g., “people under 30 working remotely” vs. “people over 30 also working remotely”)
-
Comparing across roles, departments, gender, age, tenure
-
Crossing eNPS with demographic variables
And then the right questions emerge:
-
What profiles are we combining… and which ones are we repeating without realizing it?
-
Are some groups systematically undervalued in certain competencies?
-
Does promotion really follow performance (75th percentile+) or affinity?
-
Where is diversity really located: in decision-makers or in execution?
-
Does employee experience impact groups differently?
This is where the conversation stops being aesthetic and becomes strategic.
Plan to measure and improve diversity
If you can’t measure impact, you’re not managing diversity.
You’re managing perception.
A basic approach:
-
Prioritize variables
Don’t measure everything. Choose 2–3 relevant variables for your context.
-
Measure impact, not presence
Analyze percentiles, gaps, and dispersion. Don’t stay at the average.
-
Cross variables (AND / OR)
The signal lies in intersections: not “gender,” but “gender + role + tenure.”
-
Link to decisions
Promotion, development, and team design must change based on the data.
-
Test hypotheses
Compare diverse vs. homogeneous teams.
-
Iterate
Repeat and compare each cycle.
Designing cognitive diversity (not assuming it)
Cognitive diversity is not an automatic consequence of demographic diversity. It is the result of how you combine profiles.
To design it, one way to approximate it is by analyzing which competencies stand out in each person (e.g., strategic orientation, execution, analysis, vision…) and observing how they combine within a team.
-
One profile may excel in strategy but not in execution
-
Another in execution but not in vision
-
Another in analysis but not in results orientation
That combination is what creates real cognitive diversity.
Teams where everyone thinks alike tend to converge quickly.
Teams with different strengths tend to explore more solutions.
How to frame and validate a hypothesis
We can start with observable variables like tenure and move toward competency profiles.
Example:
“Teams with greater tenure diversity (junior + senior) are more likely to concentrate profiles in the 75th+ performance percentiles.”
Expected result:
How to validate it:
If there’s no difference…
then diversity is not adding value in that context.
And that is also a valid answer.
“The collective capacity is equal to individual capacity plus diversity” and “diversity beats ability” are mathematical truths—not comforting mantras (Scott E. Page).
But like any mathematical truth, it depends on how you build the system.
You can have diversity…
or you can have impact.
The latter requires measurement.
And almost no one does it.