In parts one and two of The Antiracist Leader, I made the case that, despite all of our good intentions, the problems in our organizations still mirror the problems in our society. Specifically, the data tells us that systemic racism is thriving within our ranks, as evidenced by the diversity of our workforce, the dwindling diversity in senior leadership positions, and the rate at which underrepresented people leave our industry due to mistreatment. Refer back to parts one and two for more supporting data and links to the studies and articles that highlight these issues.

In these prior articles, I also provided tips and ideas for how any leader, regardless of their seniority or official managerial duties, can take antiracist action to affect change in their organization. For the next two articles, I’m going to take a slightly different approach.

Some things in our organizations can only be changed by those with official power and authority: things like promotion processes or workplace policies that are usually developed and run by HR leaders or senior management. That is who I’m speaking to in this third installment of The Antiracist Leader. Those of us with control over career-impacting policies have a duty to ensure we account for all the ways these policies lead to unequal outcomes for people of color, no matter how unintentional that is or how meritocratic we think we might be.

On being ‘data driven’

In the previous Antiracist Leader articles, I’ve gone to great lengths to surface data that supports my assertions. Data is an outrageously important part of our analysis and decision-making. So much of where we decide to focus our energy at work is driven by data. But in Silicon Valley and the larger tech industry, we seem to pretend that data is our only unbiased input, when in fact we know that data is hugely biased and manipulatable, especially when homogenous groups of people control the collection of data and the conclusions drawn from it.

Furthermore, a large portion of us work in organizations where data involving sensitive information like racial identity is kept under lock and key, and where conclusions often can’t be drawn because n counts are too low. For example, wide-ranging studies tell us that Black workers are being held back from advancing into leadership roles. Yet in all of the companies I’ve worked for (three of them public companies with 1,000+  engineers), I could count on two hands the number of Black or Latinx engineers who were within two steps of a management or staff level role. With such low representation, where any data anomalies can be explained by talking about each person’s individual performance and trajectory, it would almost be irresponsible to draw any real conclusions about these companies’ promotion practices based solely on advancement data. (And that’s assuming I could even get access to this data while working there as an employee, which in almost every scenario I could not).

So for the rest of this article, I’m going to stop with the studies, and I’m going to speak from my experience. This is the experience of building processes, creating or fixing policies, managing people at the intersection of varying ethnic and gender identities, and listening to the concerns of my colleagues. In particular, I’m going to highlight some common workplace policies that have, from my own observation, tended to lead to unintended, unfair outcomes for people of color. We might be able to back this up with studies done at a macro level, but mostly it’s just something I know to be true.

And these are merely examples; they might not apply to you or your company. But hopefully they will give you some food for thought, and highlight an approach you might take to root out racist policies in your workplace and counteract them with antiracist action.

A final note: I’m going to speak about a real-world example and mention a handful of others from my own working experience, and it might be tempting to look through my bio and draw conclusions about the practices of the companies I’ve worked for. In actuality, the places I’ve worked are really no different from the places you’ve worked – especially in the tech industry where it sometimes feels like every workplace policy or process is the result of someone’s attempt to replicate what they did at their previous job. My point is not to draw attention to any single company’s issues, but rather to show that systemic racism has a way of creeping into every aspect of our work, even when our good intentions have us feeling justified.

The ‘objective’ workplace policy

As one of my prior companies went through a hypergrowth phase, they suddenly found themselves inundated with requests from engineers who wanted flexibility in their working location. Some wanted to work from home a day or more per week, others wanted to move to a different location and work remotely full time. Managers were dealing with these requests on a case-by-case basis, looping in their management chain and HR business partners as necessary. Decisions lacked consistency, and some engineering orgs were more accommodating than others. Managers who lacked confidence in making a call wanted a policy to refer to; engineers who were denied the privilege of working remotely wanted to know what they had to do to get it.

One of the ways we create consistency and combat unfairness in the workplace is by removing points of subjective decision-making. If we leave it up to an individual manager to decide who from their team can work from home, someone will eventually feel slighted. Without objective policies in place, we may also be opening ourselves up to legal challenges should special privileges be granted inconsistently and unfairly.

In this example, HR recognized the need for an objective policy, and later announced this new policy to the org: all engineers at the staff level and above who had not been rated as ‘missing expectations’ in the past 12 months would be granted the privilege of working remotely if they wanted to.

Seems clear and fair right? Perhaps you already know where I’m going with this. Despite this being one of the more diverse orgs I’ve had the pleasure of working in, the junior levels were far more racially and gender diverse than the staff+ levels. In fact, this policy made it so that only a small handful of Black or female engineers (note: not Black and female engineers – there were no Black women at the staff level) qualified to work from home. In the pursuit of fairness, the team had created a policy that shut out the majority of women and people of color in the organization. To call it a sexist and racist policy would not be unjust.

Unsurprisingly, there was pushback on the level requirement. This requirement was put in place by well-meaning folk who wanted to ensure that junior engineers were set up for success and had reached a certain level of independence before releasing them into such an unstructured environment. Did I agree with this premise? For the record, no. While I think starting a career remotely is not ideal, I don’t personally view a person’s level as the best predictor of success for a remote worker, and much of my past experience supports that view. But in this case, I’d argue that my own personal rejection of this premise is irrelevant – regardless of anyone’s intentions, this policy was disproportionately exclusive to certain groups.

Dissatisfaction with this policy was not limited to the company’s marginalized groups. Many people from majority identity groups also felt slighted, and so as it goes with pretty much any company policy, people began asking for and being granted exceptions. Of course, the ability to ask for or grant an exception was not widely publicized, so you might guess who was most emboldened to request one, and who was most often granted one. The objective policy that had been put in place was no longer objective as certain people figured out how to navigate around the requirements, and now we were worse off than we were when there was no policy at all. Once the powers-that-be finally recognized the misstep, the level requirement was removed, which was (in my opinion) a small step in the right direction.

This is a great example of ‘objective’ policies gone awry. Other examples of unintentionally problematic policies I’ve seen that you might have in your org are:

• Who can have access to an executive coach;
• Who can attend conferences or trainings (or how much they can spend);
• Who is allowed to switch teams or roles;
• Who gets additional time off or other benefits and perks;
• Who is invited to attend a strategic offsite;
• Who is invited to participate in a hiring or promotion committee, or be a member of an architectural group (all of which are often nice additions to a promotion packet or résumé).

Should these even be policies? In most companies and in most cases I’d argue they should be. But the criteria that defines who benefits from them should be scrutinized for fairness.

Creating inclusive policy

You are probably thinking, ‘Okay, so then what is the right way to establish these policies?’ There’s no single answer to that, but I’ll offer a simple framework.

First, let’s establish some common pitfalls:

• Policies that only apply to a certain level of seniority will almost always wind up disproportionately excluding certain demographics, and so I discourage this wherever possible.
• Policies that favor those identified as high-potential or high-performing are better, but we must recognize that they rely on a manager’s potentially biased assessment of someone.
• Policies that lack a documented framework for requesting and granting exceptions are not worth the intranet page they are written on. Those with the strongest relationships with management or the best whisper networks will benefit more than everyone else.

The answer is not to be without policy. We should establish objective policies, but: 

• We should do the math ahead of time to determine who those policies are likely to benefit and exclude. If the criteria we’ve set excludes the vast majority of a demographic or creates a homogenous group, we need new criteria.
• To take this further, we should avoid policies that strictly favor those in senior roles. Just based off of industry data alone, we’re likely to disproportionately impact certain demographics.
• We should think aspirationally about who we want to see taking advantage of these policies, and develop criteria that does not exclude them.
• We should always provide published guidance for how to request an exception, and provide decision-makers with criteria for how exceptions will be decided.
• If we choose to base criteria on someone’s performance, we should consider how we might make that evaluation as objective as possible.

To borrow from another common example from my past experience, let’s say your company wants to put together an Architectural Review group of senior engineers who convene to discuss the organization’s long-term technical roadmap and review engineering specs of significant cross-team complexity. In this case, a level requirement is not unreasonable – you will need your more experienced people in these meetings. One easy approach might be to choose the most senior engineer from each relevant team and call it a day. But let’s say you do that, and you find out that your group is looking overwhelmingly homogenous. Now perhaps you come to the realization that a homogenous group is unlikely to net you the innovative results you surely want (I said I wasn’t going to link studies so you can Google this one if you need to). One approach that a prior company took was to:

• Set a top-level goal for the diversity of this group (not a specific quota, but an explicit, overall goal that everyone could buy into);
• By default, include everyone in the top two engineering levels (this was a rather small number of engineers);
• Frame up the goals and responsibilities of this group in a request to Engineering Directors (including stating the goal of creating a diverse group), then ask for nominees who:
  • had expertise in some of our more critical systems, and
  • showed growth potential that made us feel they would benefit from access to this group;
• Rotate all nominated engineers on a 6–12 month basis.

This approach allowed us to staff this group with the senior folks we needed to have in the room, while ensuring the group would be more diverse than our most senior levels. It didn’t absolve us of our responsibility to ensure our most senior levels were more diverse, but it did provide an avenue towards accomplishing that goal. We were able to give a handful of promising, underrepresented engineers more access to senior peers, make them more visible to senior leadership, and give them space to demonstrate wider impact than they typically were able. Creating a rotation meant that, had the group endured (okay, no story is perfect), we could offer this opportunity to more folks over time.

As I mentioned, these things are tough to get right, and there’s no silver bullet. But with intention, we can identify when we’re creating policies that are unnecessarily exclusionary and/or require political know-how in order to circumvent the policy’s criteria.

What’s next

A few times throughout this article I referred to company processes, and specifically mentioned policies where the criteria is based on an individual’s performance. Performance evaluation is of course a thing that can be fraught with bias. While I won’t claim to be able to solve that problem for you, in the next and final installment of The Antiracist Leader, I will cover a framework that has helped me surface and reduce bias in organizational processes, such as performance evaluations and promotions. Stay tuned!