Career DEI women in computer science

From Visibility to Representation – Rethinking DEI

When I talk about Diversity, Equity and Inclusion (DEI), I’m typically coming at it from an angle of systematic change. The purpose of DEI, as I see it, is to dismantle a rigged system and move to something more equitable. This is why the concept of “no politics at work” is seen as antithetical to effective DEI, because what does a person do when their entire existence has been politicised?

The frustrating thing about DEI, is that often when organisations talk about DEI what they mean is the performative type of DEI. The appearance of progress, without the challenge of systematic change. The percentage points that can be shared externally, like there’s been progress, when the balance of power remains the same. The updates that start and end at the company website, and leave out the hiring process, the promotion process, and anything else that might threaten the status quo.

This is the trap of Diversity as Performance Art, or (because it so often focuses exclusively on white women) “Corporate Feminism”. Plenty of us have fallen into it, spent time on it, only to later look at the state of things and wonder what it accomplished.

Personally, I quit Corporate Feminism some time ago. Not quite entirely, I have dabbled since, embraced the odd moment like a not-entirely-ex smoker sneaking a cigarette, feeling grimy afterwards and remembering why I chose not to do that anymore.

It is hard, I think, to build a model of what DEI really is in tech, amidst all the noise, and “systematic change” can be overwhelming as a concept. Here, I’ll try to lay out some component parts of DEI and explore the impact and limitations they have.

  • Visibility vs. representation
  • Education vs. advocacy
  • Mentorship vs. Sponsorship
  • Individual vs. Environment
  • Status Quo vs. Structural change

Visibility vs. Representation

Visibility often seems like the most fraught aspect of DEI. On the one hand, we hear “you can’t be what you can’t see” and many firsts or onlys are driven by the idea that they can be the role model they lacked to others who come after them.

On the other hand, the concept of exposure, and the expression “people die of exposure”. The costs of visibility in terms of harassment and resentment. The expectation of visibility which is at best the unpaid second shift, and at worse, free work “for the good of the community“.

At the crux of the issue of visibility is the distinction between visibility and representation. Visibility is a face on the corporate website, or additional interviewing load, to make the team look “more diverse”. Representation is the under-indexed person in the hiring manager role. Visibility is the speaking slot at the sponsored recruiting event. Representation is the invited talk at the industry event based on actual expertise. Visibility is the pride flag on the Twitter profile. Representation is in the people who actually work there and the policies that actually support them.

At the core of it, visibility is some form of currency, that occasionally exchanges for some goodwill, but rarely anything more. Representation is a form of recognition; representation is visibility for some reason that ties to the actual work you’re compensated for. The first shift of your actual job, not the second, thankless, one.

I’ve noticed when I feel tokenized, it’s typically been that I was visible but not representative. The speaking slot where I thought I was invited because of my expertise and previous talks, only to doubt that at some point during the negotiation, or worse, after I arrive. The recruiting email that sells the diversity of the organization and the list of roles… that make me suspect the recruiter never even looked at my resume.

As a result, I’m careful about what I agree to and what I don’t. I think every under-indexed person has to find their own balance there.

Takeaway: Think about your balance of visibility vs. representation. What is valued? What are you getting out of the things you agree to?

Education vs. Advocacy

Education is information. Advocacy is pushing for change based on understanding of systematic bias.

In essence, advocacy is an educated person, with the power and will to make change.

The trouble with education is that it often focuses on 101 education, and someone who needs 101 education has made a choice to be ignorant about the world. It also doesn’t necessarily create the will for change. When meritocracy is adopted as a value, there is more bias, there’s also the potential for a shallow understanding to create harm.

Needing to provide 101 education is also a distraction, time sink, and often starts with needing to justify someone’s existence. There is a cognitive cost to explaining things like “no, women don’t just ‘not like coding’, that is not a well founded statement and has been refuted many times over by [list of sources]”. That kind of comment is annoying, and tiring, but the base level comments about people of colour and trans people (trans women in particular) are so much worse. There is the idea that personal stories can change minds in a lasting way, but that was only 5-7% in political canvassing, so as an activity it has a high disappointment rate [more in this podcast] – as such it’s probably more suited for political canvassing than for people you need to interact with regularly.

The trouble with advocacy, is that women are perceived negatively when they are seen as self-advocating, but more positively when advocating for others (although not as much as men) [see the book Women Don’t Ask, and this HBR article Women and Minorities Are Penalized for Promoting Diversity]. The trap of education and advocacy for under-indexed folk is that in education you may waste your time, and in advocacy you may in fact set yourself back.

As such, the only worthwhile use of education is the level that creates advocacy. Education in the absence of both the power and will to make change is meaningless. This means: nuanced education of what actually works to people with power and will to change.

Takeaway: Be cognizant of the traps of education and advocacy, and critical about where you spend your time. Often this means avoiding 101 education, and focusing your energy on those with the power and will to make change.

Mentorship vs. Sponsorship

Mentorship is giving someone advice, like pairing with a junior developer or helping them find resources. Sponsorship is more active help to change someone’s situation, such as connecting them to an opportunity or advocating for them.

Mentoring programs have been very popular, but not particularly effective. This is because mentoring is a Ponzi scheme designed to distract under-indexed people from actual advancement. Or, less cynically, because the problem isn’t the individual but rather the system.

Sponsorship is in fact the biggest lever for an individual. The challenge of prove it again, of judging some people on past performance and others on potential, is that decisions on opportunities and advancement are rife with bias. Sponsorship programs, especially those that hold sponsors accountable for supporting their sponsees advancement, attempt to correct for that bias and create more opportunities for success [see: Why Men Still Get More Promotions Than Women].

Mentoring programs often pair people up with another person “like them”, and whilst it’s nice to build connections with people who share similar experiences, ultimately it often creates additional work for both parties. The mentor is normally further away from the person they are mentoring, and lacks the power to truly advocate for them. In a sponsorship relationship, the sponsor has influence over the sponsee’s work, can give them more feedback, connect them to opportunities, and shape their career.

Sometimes sponsorship programs are formalized, but they don’t need to be formalized. All that is required is that people who have the power to do so look for ways to advance people more equitably. Sometimes it’s a one time thing, and sometimes it’s an ongoing thing [see: 5 Things Allies Can Do to Sponsor Coworkers from Underrepresented Groups, Don’t Just Mentor Women and People of Color. Sponsor Them].

Personally, I find sponsorship is one of the most rewarding activities. Nothing is better than seeing someone you helped get into a position where they can pay it forward to someone else. Sponsorship is a slow but compounding process, but ultimately one of the most powerful ways for individuals to escape traps of systemic inequity.

Takeaway: Quit the mentoring Ponzi scheme, seek out sponsors instead (and be a sponsor yourself, where you can! Sponsorship can start small).

Individual vs. Environment

In Stop Telling Women to Smile [Amazon], the author Tatyana Fazlalizadeh talks about her experience in a street artist collective. She was on the receiving end of street harassment whilst doing the same work, day in day out, and how the others in the collective – all men – were able to get on with what they were doing. She talks about worrying about her clothing when it was hot, whether wearing shorts would increase the verbal abuse, whilst men would work shirtless and not deal with anything, never say anything about the abuse that was hurled at her.

This story is so illustrative of the tax of the bad environment. The cost of distraction, the unfairness. The environment is a significant piece of what people mean when they talk about “working twice as hard for half as much”. Some of that extra work is work, and some of that extra work it just… getting to do the work.

This is why an inclusive environment, one that attempts to correct for systematic bias and ensure that a broader range of humanity can be successful can be so impactful for improved DEI outcomes. It provides people who often do not get such a thing a space where they can focus on the work at hand, without the tax of hostility, or thoughtless ambivalence. Individuals get the opportunity to thrive in an environment, rather than despite the environment. It is great for people in it, and it can create a model that encourages change in other parts of the org – or at least discredit the pipeline arguments that say such diversity isn’t even possible.

The limit of the environment is what is around you. Even if you create an environment where under-indexed folk can thrive, they will have to interact with other parts of the organization, and you won’t be able to control what they encounter there. When you interact with shared process (such as hiring, compensation, etc), these can be some of the biggest detriments to your effectiveness in creating change. It can be exhausting to do that transformation and provide that abstraction, but how long can the principles you have laid down last if you leave?

Takeaway: Seek out environments where you can thrive, and as you build influence look to what you can do to make an better environment for those who come after you. As you do this, don’t forget to pay attention to where the boundaries are – both of the change you can create in an organization, and what you can personally bridge.

Status Quo vs. Structural change

Processes create and reinforce a status quo, and they are often the place that needs to be changed in order to make lasting improvements

When we design process, we are biased to design process we would be successful in. This is why process updates are inevitable part of effectively improving diversity outcomes.

I have repeatedly overhauled hiring processes to improve diversity outcomes. The key thing as always been to get very clear about what success in role looks like, evaluate that carefully and systematically and as efficiently as you can, actively address signs of bias (e.g. the words “culture fit” are lazy, so at least interrogate what people mean when they use them).

I have not (yet) overhauled a promotion process, but I expect that similar thinking applies. Whenever people claim they “know it when I see it”, we need to interrogate what “it” is. Once that is articulated free from aspects of bias (put things like “good company on the golf course” in the bin where they belong), you have the start of a meaningful evaluation which you can reason about and improve. Are there impactful behaviours being left out because no one did them before? Maybe those need to be added. Are some of those things particular ways of working and the same outcomes can be achieved differently? Replace the activities with the outcomes they drive.

When we accept we live in a white supremecist patriarchy, we can see that white supremecist patriarchy is everywhere. It is in the media, in beauty standards, in relationship expectations, in politics and religion and art. It is in ourselves and our judgements and beliefs – we can try to consciously dismantle it, but we can never be confident it is truly gone.

As such it is inevitably in every process we create, and that makes constantly critiquing process with respect to outcomes the way to – albeit incrementally – start to dismantle it.

Process is a powerful mechanism for upholding the status quo. Consider VC funds. Encoded in their process – and many of them are pretty open with that – are pattern matches. When Paul Graham famously said “I can be fooled by anyone who looks like Mark Zuckerberg”, what he told us was that pattern matching is such a key part of his investment decision process that it could override other, presumably more rational, parts of the assessment.

I believe process as a tool is most powerful when it can affect redistribution of power (and money). This is why such comments from VCs like this get so much heat, because who gets funded now is huge part of who will have power 10-20 years from now. The longer inequity in hiring and promotion processes is perpetuated, the more it will cement homogeneity in the leadership of an organization.

Takeaway: Process is a powerful mechanism for change, and making it less biased can also make it more effective for everyone. If you can frame things around organizational effectiveness, you may even be able to make change without rehashing tedious DEI 101 arguments along the way.

Now What?

There’s an excellent book about stereotype threat, Whistling Vivaldi [Amazon] by Claude Steele. One of the examples that most impacted my thinking is the power of self identification at start of the math test, and how it impacts people’s results – when asked to self-identify gender, women score lower, when asked to self-identify race, Black people score lower. This demonstrates the insidiousness of systematic inequity; that social programming is so strong, it’s sufficient to remind people of it. It is not necessary to “do” anything to perpetuate systemic inequity, the absence of care and effort alone is enough. The action is insufficient to judge whether it perpetuates bias or whether it addresses it – when asking people to self identify you ask them to all do the same thing. But it has impact on some, and not on others.

It has been 10 years since Ellen Pao sued Kleiner Perkins for gender discrimination. Nearly 10 years since Tracy Chou asked where the numbers are… what has changed in that time?

The conversation has expanded, somewhat. We talk more about racial equity and transgender rights.

Worker agency has improved, at least in some parts of the world. People set higher standards for environment, and process improvements have been made across much of the industry. What started as a differentiator is now more normal.

But representation still lags visibility. Sponsorship lags mentorship. Environment is hit-or-miss. Bias is still encoded in processes. Companies still focus on the things that are comfortable over the things that have more impact to change the status quo.

Personally I go back and forth between hope and cynicism. I get angry and resent being exploited for visibility, feel that my efforts have been wasted because too many people in power do not really care, are not really willing to change anything hard. And then I have moments where I can recognize my own impact, where I see the sponsee pay it forward, see the person thrive in an environment they never got to experience before, see improved outcomes from more equitable process. When someone thanks me for something that I wish everyone got to experience, but at least this person did, at least I could do that for them.

I hold on to those moments, because those are what I come back to when I get discouraged. Those are what push me to keep on trying. And this framework helps me decide what to spend my energy on, and what to let go.

Thanks so much to Karen, Heidi, Trisha, KT, Elle and Camille for their feedback and validation.