Focus on Impact Instead of Outcome for DEI
Defining success based on impact can steer DEI initiatives in the right direction.
Background
In May of last year (2024), I wrote an article on why DEI initiatives fail, citing poor data definitions and measurements, and the overall lack of data fluency skills by DEI practitioners. That article has since garnered over 32K views and is trending at #1 on Google search. All this was due to Trump’s recent roll-back of DEI initiatives across the US, triggering a massive global inquiry into the “why” and what” of DEI, a term many outside of the US or western hemisphere may not be familiar with. (I operate in Singapore.)
Because of the viewership of this said article, I’ve been invited to an interview with the BBC (which I turned down), and I’ve been invited to discuss how data and DEI intersect on a podcast and possibly to a US-based conference. And so I dedicate my 80th article as a follow-up to why DEI initiatives fail, and how one might approach it from a data-centred perspective.
(I write a weekly series of articles where I call out bad thinking and bad practices in data analytics / data science which you can find here.)
Ground Truth
Let me first qualify that I support the need for diversity, equity, and inclusion in society and in workforces, and have pushed for those things in quiet ways during my leadership roles. I am tri-racial and so have first-hand knowledge about bigotry and exclusion. Nonetheless, I also understand the pushback on DEI. Sustained affirmative action, when taken as an emotional response to social injustices, can create more inequity over the long term; I’ve seen the devastating effect it has had on my neighbouring Malaysia.
Interestingly, a number of my pro-DEI HR C-suite friends welcomed Trump’s roll-back of DEI. They believed in the principles of DEI, but not how it has been operationalised. A large chunk of it comes down to how success has been defined, and hence measured (if at all).
Outcome vs Impact
The principles of DEI were articulated to achieve a better world for ALL, and that has to include organisations and society as well, and not just the minority or marginalised groups that DEI is meant to uplift.
Creating good metrics starts with understanding and applying the framework of Input → Activity → Output → Outcome → Impact, something I first wrote about in my article The Problem with Dashboards. Many DEI initiatives go unmeasured, but when they do get measured, they are measured for their output and outcomes, not their Impact. For example, a DEI initiative might measure the number and proportion of women hired and promoted as part of gender diversity and equity. Or DEI initiative might measure the reduction in pay differences between a minority versus majority group. These are all output and/or outcome measures. But they are not impact measures. And when impact goes unmeasured, people make up narratives through their own observed attributions which are often (unintentionally) biased and incorrect.
Writing about these measurements reminds me of the time when I was first starting out as a country chief analytics officer (CAO) for Citibank. This was a time before the lexicon of data science, and the data analytics function was nascent. That means there was a lot of time and effort spent on capacity and capabilities building. A fellow country CAO gave me this invaluable advice: “We need to work on the function, but don’t forget we also need to work for the function.” Working on the function is what most DEI initiatives are doing, showing that they can help their in-group members achieve their target outcomes. It’s like measuring how many campaigns the data analytics team could execute and/or automate, or how many sophisticated predictive models it can deliver. But working for the function is about making the function successful in the eyes of the organisation and its decision-making stakeholders. And that means delivering incremental P&L, highlighting blind spots and unanticipated risks, and bringing efficiencies to information processing and dissemination. The equivalence to DEI initiatives is to measure success from the perspective of the organisation, the customers, and the community.
DEI Impact
Revenue, cost, innovation, and perception — these are the 4 important impact that DEI needs to be measured on. For example, DEI leaders need to work with their data analysts or HR Analytics team to find answers to such questions as:
- Does representation (e.g. increasing the mix of minority hires) lead to increased penetration into new markets and new revenues for the organisation? And if so, what is the desired mix to achieve and sustain?
- Does supporting sabbatical with the “right-to-return” (e.g. holding an open position for women to take significant time off for pregnancies) lead to a reduction in replacement cost and better productivity over time?
- Does cultural and experience-based diversity lead to an increased in the number of innovative projects in different departments? And if so, how best can we tune the nature of this diversity to further optimise desired innovation outcomes?
- Do positive employee experiences from DEI initiatives make it easier to attract desirable talent for the organisation, leading to an overall reduction in hiring (i.e. seek & search) cost? (An example here is the recent public sharing of DEI student experiences in Princeton by the university’s president.)
- How much goodwill (e.g. business loyalty, social media support and compliments) is generated from the customers that we serve and the community in which we operate because of our DEI initiatives?
Conclusion
Yes, there is some hard work to attribute revenue, cost, innovation, and perception impact to DEI initiatives. But we shouldn’t shy away from it. Instead, DEI leaders need to lean into getting support from their data analytics community. Simply measuring the outcome of DEI initiatives is like patting oneself on the back. Measuring the organisational, customer and community impact of DEI is about being a valuable partner to your stakeholders.
And DEI leaders shouldn’t stop at just measuring impact, but also re-think how to go beyond the archetypes of DEI in a way that raises the game for everyone. For example, there are broader ways to think about diversity. I personally like to think of diversity as hybridisation. Evolution tells us that hybrids are more resilient. We should therefore expect the same to apply to organisational evolution. Diversity isn’t just about ethnicity or gender or sexual orientation. There is also thinking diversity — see the book Collaborative Intelligence: Thinking With People Who Think Differently by Dawna Markova and Angie McArthur, which shows how we can unlock more performance potential by recognising friction in the way we “tune in” to each other.
