We had a lot of beliefs about who worked at Table XI, but not a lot of data. We had never done a diversity survey to see how people identified, in part because we weren’t a large organization and in part because we felt like we already knew the answer. Which is proof we fall into the same trap we’re constantly helping our partners avoid.
Forget what you think you know and ask questions of the people you’re trying to understand.
To pull together the right survey questions, we had to think about what we were trying to learn, where it would ultimately be shared and most importantly, how we could make everyone feel comfortable sharing information about how they identify. It took us about half a year to figure out, but what we have now is a repeatable and representative process we can use to put real goals behind our diversity efforts going forward.
Defining the questions a diversity audit can answer
While we may have a lot of experience doing user research, the goals of a diversity survey are different. It aims to put people into quantitative categories — the questions we’d write for a qualitative user interview don’t apply.
Still, we knew how to approach the problem. The first thing to do is understand what you’re planning to do with the data you collect. In this case, it wasn’t about getting everyone into neat little boxes or reporting out to shareholders. It was about understanding the trends in our workforce so we could intentionally correct for overrepresentation.
With those goals in mind, we looked for diversity and inclusion reports that took the same approach, looking at the questions those companies relied on for answers. For this, diversityreports.org was especially helpful.
Starting our workplace diversity survey with a survey about the survey
That gave us a sense of where we wanted to go, but we still needed to figure out how to get there — and do it in a way that felt safe and respectful for everyone. We took what we learned from our research, synthesized it, then turned around and asked our team what they’d like to learn and answer.
We started off with the fundamental principles that the survey would be anonymous — or as anonymous as you can make it in a 50-person company — and that anyone could opt out of any question they wanted, as well as the survey as a whole.
Surveying about a survey may seem redundant, okay it is redundant, but it was also necessary. It clarified what we were looking for and brought valuable perspectives into the project from the very beginning, improving the outcomes.
Creating a quantitative diversity survey that values self-determination as much as structured data
The goal of an employee diversity survey is to make our team more representative. If we reduce people in the process, that’s creating a worse outcome. We want people to feel safe and comfortable identifying who they are without fear of being singled out. That meant embracing identity as a shifting and overlapping thing, while still working toward hard numbers we could understand and improve on.
Whenever possible, we leaned on self-reporting instead of forcing static data. Even just talking about “normalizing” data has all kinds of connotations we wanted to avoid. When it comes to data about who people are, it’s messy and fungible and dynamic and hella hard to work with and that’s all okay.
Protecting anonymity and encouraging participation to include everyone
It’s hard for any one of us to say that we don’t know the people we work with. Table XI is a 50-person team, and a close 50-person team. It takes a lot of effort to see a few data points and not immediately connect them to the full person you know and respect.
Still, we wanted to preserve anonymity as much as possible. The point of the diversity report isn’t to single anyone out — it’s to look at where we are as a group and identify where we want to go. Any time it was possible to anonymize groups under five people, we did that, while being careful not to constantly throw less-represented groups into an “other” category that didn’t fit anyone.
When the diversity survey went out as a Google form, we more or less relied on interest and the honor system. We asked people not to answer more than once, and we encouraged people to participate without making anything mandatory. Ultimately we had 80% of owners, contractors and staff respond, a bar we feel is pretty good for our first attempt.
We did miss some folks, including at least one person who was surprised not to be represented in the data. In the future, we’ll do a better job making clear that this survey is the only source of diversity data within the company. We’re not collecting demographic data anywhere else.
Translating our data into a diversity report we can build on
When it was time to translate the results into infographics that would inform our actions as a company, we had to call in reinforcements. Preserving anonymity while presenting meaningful narratives gave us a thin sliver for error. Fortunately we were able to call on our friends at Ethos Talent for support.
Our first annual diversity report is a snapshot of Table XI in this moment. It doesn’t represent who we’ll be for the entirety of 2021, or who we were for all of 2020. It doesn’t represent how any one person is going to define themselves year over year. What it does represent is our willingness to look honestly at who we are and take actual steps to become who we want to be.
We’ll be using the survey data to answer all kinds of questions, from when to go back to an office to how to conduct surveys like this going forward. For now, it gives us the floor for our inclusion and diversity hiring initiatives. Over the next few years, we’ll be trying to decrease overrepresentation by 10%, while reevaluating to see what diversity goals make the most sense with next year’s data.
If you’re interested in seeing how it goes, please check out our newly launched DEIB page. We’ll be updating it with new data and new efforts as we go along.