A New Framework for DEI
- ️Sun Feb 23 2025
With escalating attacks on corporate diversity, equity, and inclusion (DEI) efforts, many private-sector organizations are searching for an approach that preserves effective, popular initiatives while insulating them from external attacks.
For strategist Lily Zheng, the solution is not wholesale retreat, but holding DEI programs to the highest standard of effectiveness using a framework they call FAIR: fairness, access, inclusion, and representation. “FAIR is built on four principles that make DEI work effective, namely being outcome centered, being focused on systems, being driven by coalitions, being communicated in win-wins,” they explain.
They contend that many programs currently being rolled back—such as participation in external surveys and public statements on a company’s website—fall short of this standard. We spoke to Zheng to learn more about how these kinds of programs have failed in the past and their recommendations for designing DEI strategy that will stick going forward. Here are excerpts from our conversation, edited for length and clarity:
You have argued that much of mainstream DEI was falling short, even before this current wave of attacks. Why is that?
There was a study that came out showing that between 1990 and 2015, which is 25 years, the rates of anti-Black hiring discrimination did not change by a single percentage point. There’s lots of similar stats on the glass ceiling for women on promotion rates being low, the bamboo ceiling for Asian people, disability-related discrimination rates. By all of these measures, we’ve seen that workplaces haven’t necessarily been making good progress.
All these problems still existed, but companies tried to solve them in ways that were just ineffective. You don’t solve racism by having a book club. You solve racism by fixing your hiring process, by training your managers, by ensuring that there are consequences for harmful behavior, and by building respectful norms in your workplace. This is what the work looks like and companies were not investing in it. Unfortunately, even in 2025, this continues to be the case in the vast majority of workplaces. This is the big challenge that DEI practitioners have had for decades at this point, even if backlash had not become a challenge. Now not only are we navigating backlash, we are navigating this legacy of ineffective responses to solving these deeply rooted problems.
Are there ways to make DEI work more impactful while shielding it against this new wave of attacks?
I mentioned that study done on anti-Black hiring discrimination. The same study found that the companies that had the least hiring discrimination all had one mundane thing in common: a centralized, standardized hiring process that used scoring rubrics and hiring panels. That made it so there was some rigor and a standardized process so that a single hiring manager could not hire someone willy-nilly because they liked the way they look.
So if you ask me how to mitigate hiring discrimination, I’m not going to recommend you sit everyone down for three hours to train them on every single racial bias that exists. I’ll just say standardize your hiring process. People of color or women or LGBTQ+ folks, these are not the only people that are impacted by bias. Basically, standardizing your process is going to help pretty much everyone. It ensures that we don’t act on these biases that we might hold about who a smart person is or a skilled person is.
There are lots of other common-sense interventions like this across the workplace that we don’t even need to call DEI work. This is just how to run an effective, responsible, healthy workplace that people trust and that people feel respected in.
You focus on measuring outcomes and building accountability in your work. How do you recommend organizations set these systems up, and what should they be measuring?
One of the things that I have helped companies do for the last eight years is assess their workforce sentiment on their policies and processes. We do audits and culture surveys to assess people’s access to decision making, their access to opportunity, their perceptions of psychological safety or physical safety. We analyze all these data by demographics and organizational factors like level, tenure, and department.
Data allow us to see which problems are real, which problems do not exist, which things don’t need solving, and allow us to allocate our resources to actually fix the things that need fixing and ensure everyone has good outcomes. I’ve worked with organizations where women have better outcomes than men. In that situation, how can we make this organization work better for people of all genders? I’ve worked in organizations where Ivy League grads ran a top clique—how can we make sure that talent, regardless of where you went to school, can shine?
Also, data are the first step to accountability. If a leader says, ‘I want to end discrimination,’ then of course you need to measure discrimination. You need the leader to do something about it, and then you need to measure discrimination again to see if it went down. This is the building block of any kind of effective change-management organizational-intervention approach. You measure, you do your experiment, you measure again. The more data we collect, and the more effectively we collect it and use it, the better we can not only make change, but also visualize that change and hold ourselves accountable.
What has been the impact of the recent rollbacks? It seems like there has been a mix of programs affected, from the window dressing of lunch-and-learns and heritage month messages to more concrete actions like supplier diversity pledges and diverse hiring slate policies…
I really appreciate you pointing out that not all rollbacks are created equal. There are some which are clearly performative—just as performative as the performative DEI efforts were. For example, you can get a perfect score on the Human Rghts Campaign survey and still be a terrible workplace for LGBTQ+ people. It’s just a signaling effort. In the same way, companies saying they’re not doing the survey anymore doesn’t tell me anything, but some folks on the right will see this as a political win.
But there are other things that are going to have real impacts. For example, the companies that are rolling back on their commitments to ensure broad representation within candidate pools before they hire. These efforts meaningfully increase the diversity of workplaces in perfectly legal ways because they’re just trying to extend the talent pool rather than setting quotas. These things are legal, they’re effective, they’re very well appreciated. I can’t see any possible benefit except to signal to certain groups that you’re willing to inflict harm upon marginalized communities. Or let’s say you’re funding a nonprofit or an institution dedicated to racial justice. It does send signals, but it also has real tangible impacts on everyone who works there.
Then, importantly, that’s going to really, really hurt morale for everyone within their workplace. If these things go away, there’s a very good chance that workers in those companies will feel a lower sense of belonging, lower wellbeing, and lower productivity. They’re likely to see these workers leave because there are lots of workplaces that treat workers of all races and all genders well, so we’re going to see a brain drain away from these workplaces that have signaled that they would much rather pander to authoritarians than build workplaces for everyone.
Read Zheng’s recent article about the FAIR framework in Harvard Business Review.
Charter Pro members can read a full transcript of our conversation on our website.