This post is Part 2 in a series of the ongoing DEIB journey that we are undertaking as a venture capital firm. Make sure to read Part 1 that outlines the uncomfortable conversations we had as a firm that propelled us to take further action and Part 3 in the coming weeks to learn more about the lessons learned from our 2021 DEIB Survey.
As MATH started down the DEIB journey, the one thing EVERYONE on the team agreed on is that we needed to measure our progress over time. We decided we needed a survey tool, which led to a series of questions. What type of survey would be most appropriate? More importantly, would our portfolio companies see the value in participating? What level in the organization should we measure? How do we craft a survey that actually gives actionable insights?
To begin, I researched as much as possible on best practices in the DEIB survey space and talked to many under-represented founders, organizations working on these issues (including All Raise, BLCK VC, LatinxVC, PEWIN, and Chicago Blend) as well as numerous potential service providers. Our intention was to uncover what other firms were doing well and build on their lessons learned.
Instead, we found that while almost everyone agrees measurement is important - there are really no standards for Venture Capital firms trying to administer a survey like this to portfolio companies. Most best-in-class DEIB surveys were not designed to be used this way. The two surveys that we found that were specifically designed for use by investors did not get at the richness of the data and insights we hoped to uncover. Ultimately, we partnered with Peoplism, a DEIB consulting firm, and with their help, we uncovered a few themes that we refined and made our own.
In an effort to pay it forward, below are the top five elements we feel are most important when developing a DEIB survey.
1. Use Self-Identified, Unattributable Data
The most shocking thing I found when I was researching what other Venture Capital firms and organizations were doing in the space was that they were creating reports based on an assumption of how individuals identified and then tracked that data by name. This was especially true for some funds that propose to invest only in a particular demographic profile.
To be fair, early in the MATH journey we had implemented this for founders. Diversity has always been important to us and we wanted a way to track how we were doing. We no longer do this because countless conversations with founders taught us about how they felt assigned labels contributed to a biased experience. It was also under the advice of every expert we talked to - there is simply no way to know which group or groups someone feels like they identify with and choosing for them can be worse than doing nothing. When you know better, you do better and so we agreed at MATH we would only utilize self-reported data going forward.
Getting unattributable or anonymized data presented other challenges. First, we debated whether MATH should know results by portfolio company. We did not want any of our companies to feel called out - this is about moving forward together. We decided for the data to be useful and actionable to our participating portfolio companies, we needed a report for each company in addition to a portfolio wide report. Second, at the individual level, we wanted each participant to feel safe in being completely honest. Our companies are different sizes and cultures and we felt it was important to ensure no one could be personally identified in the results. Thus, at smaller companies some demographic data had to be combined, trading anonymization for insights by group. We felt these choices created the best balance between providing useful insights and a safe place for participants to be 100% honest. In order to accomplish this, no one at MATH had access to the individual data. Instead, Peoplism administered the survey and analyzed the results for us.
2. Get the D Right
Early in this journey, I had the sinking feeling that looking at diversity data in the wrong way could actually hurt our DEIB efforts. I don’t like the way most surveys report on this type of data, continually separating and dividing humans into groups sensationalizing our differences. I also learned the importance of intersectionality. Many templates make you check one box. Liz Kofman-Burns, Co-Founder from Peoplism put it like this, “Great DEIB surveys take an informed approach to collecting demographic data. Diversity is complex and intersectional, and no one wants to make individuals feel reduced to a statistic.
Diversity can come in all shapes and sizes far beyond race and gender. For example, do you want to collect information about religious or political party affiliations? How about socio-economic status which is incredibly important and difficult to measure - usually collected by asking about the education level of the person’s parent/legal guardian or whether they are a first generation college graduate? Maybe you want to capture information between remote and in-office workers. What about caregivers and non-caregivers?
There must be a balance and it’s important to take into consideration your company size, objectives for collecting the data, DEIB strategy maturity, relevant benchmarks and resources for analysis.
3. Go Beyond the D
I see many investors and companies collecting only diversity/demographic data and not focusing on EIB (Equity, Inclusion and Belonging). We get it. The D feels solid, its factual and numbers based. But for MATH, it's the equivalent of receiving a financial model from a founder that shows revenue only but none of the drivers of the business – it doesn’t really tell us much.
The magic happens when you take a look at survey results about inclusion and see noticeable differences in the way different demographic groups answer the questions. For example, if you have a culture problem and all employees experience that problem similarly, you have a clear set of issues. If you have a culture problem that only one demographic group experiences, that is an insight that points to a different set of solutions. It tells you that something is not only off, but off in a way that could be caused by or contributing to bias and divisiveness in your company.
4. Craft Actionable Questions
The most important thing we learned from working with the team at Peoplism was the importance of creating survey questions that are actionable. Unfortunately, many common DEIB surveys don’t have the most actionable questions. Actionable questions inform what action you need to take if you see disparities in the results.
Maybe you learn, for example, that Black employees at your company are less likely to say they have a voice. How does this finding drive toward a solution? You first have to really understand what is going wrong. In this case, is it that there are no Black executives at your company, so Black employees don’t feel heard at the highest levels? Is it that managers are systematically excluding Black employees on their teams from key projects or decisions? Are peers interrupting Black colleagues in meetings? You don’t really know from a question like that.
In contrast, actionable questions inform steps to take to improve. An example of an actionable DEIB question in Peoplism’s DEIB survey is: “My manager actively seeks out my opinion on important matters.” Suppose your found that employees from underrepresented groups are less likely to agree with this question than employees from majority groups. Your actions would be to train managers on the importance of involving the whole team in decision making, education on the biases that may prevent them from doing that, and elicit the behaviors managers can use to solicit more input from everyone who reports to them. Actionable questions drive results.
5. Complete After-Survey Actions
This is an obvious step but worth noting - if you are going to survey, you must take action. We decided that we were going to offer our portfolio companies two different training workshops in 2022 based on our 2021 survey results. The first is focused on sourcing for diverse talent and the second is focused on ensuring bias is taken out of the interview process. In addition, each company that participated in our survey had the opportunity to debrief with Danielle Callendar, Senior Consultant with Peoplism to help set individual company level actions based on their survey results.
The final word here is that we were very open with our portfolio companies that doing this work is a journey we are on with them - and these resources are meant to be like everything else we do at MATH - help them succeed with greater certainty.
We would love to learn more about effective DEIB initiatives that you have been a part of as a founder or a VC. Sound off in the comments.
Follow @DanaWright_VC and @MATH_V_P to learn more about our work and process.
Post to Come:
Five Key Lessons Learned from our 2021 DEIB Survey (Part 3 of 3)
