As for any strategic ambition, monitoring the progress of diversity requires data. But beware of misleading statistics! More so than for other topics, a quantitative approach can not only produce an erroneous vision of the company’s real situation, but it can also undermine the credibility of its commitments in terms of DEI (Diversity, Equity and Inclusion).

Two traps more specifically threaten this type of statistics:

? The developed averages often hide pockets of over-concentration: such profile type is over-represented for a given job expertise or hierarchical level, but almost absent elsewhere. This is why Nike, IBM or Salesforce demand that demographic data always be presented at the finest possible granular level, at least differentiating the different hierarchical levels and the technical or non-technical positions.

? Data are only meaningful in their context. The objectives and progress in terms of DEI will not mean the same thing for an entity located in a zone that is itself mixed or in a zone that is marked by the massive prevalence of a certain population group. This reminder might seem obvious, but processing data that are aggregated at headquarter level might rapidly make you forget the strong influence of local anchoring.

Source: Improve Your Diversity Measurement for Better Outcomes, Derek R. Avery, Enrica N. Ruggs, Larissa R. Garcia, Horatio D. Traylor, Noelle London, MIT Sloan Management Review, November 2022.