Clover’s 2018 Diversity Report: Progressing Toward Best Practices
Last year, Clover Health began a commitment to externally report some of our demographic data, a commitment we are excited to continue this year. We believe doing so not only furthers our promise to be transparent, it also adds a layer of accountability to our progress over time.
Many companies talk about the business benefits of diversity: different perspectives coming together to create better products and results for their customers. Here at Clover, this can literally be a matter of life and death. Healthcare touches everyone, so we need employees with different work experiences, backgrounds, and points of view to effectively solve the complex problems faced in Medicare.
“At Clover we strive to use data to inform all decisions, and Diversity and Inclusion are no different. Diversity is an emergent property of our focus on equity and fairness, and by looking carefully at our data, demographics and sentiment we can learn what we’re doing well, what we’re not and how to improve. That’s the Clover way!” Andrew Toy, CTO
But we also understand that diversity without inclusion and equity is harmful to all. To hire people from diverse backgrounds and not provide them with equitable opportunities and an inclusive culture perpetuates systematic inequalities in broader society which we are determined to address at Clover. Our approach is simply that inclusion and equity come first. This means that while we continue to invest in diversifying our recruiting pipeline, we are collecting even more data on employee experience than on demographic representation. This is not work that I, as the company’s Diversity & Inclusion Manager, can do in a silo. This is work that everyone does together and is a hard, long journey, but it is one we’ve pledged to continue.
Where we’ve started:
Our approach to inclusion continues to develop at multiple levels of the organization. Our most significant initiatives, starting very early in the company’s history, are the Diversity & Inclusion Working Group and Community Engagement Group. The first focuses on a range of Diversity & Inclusion initiatives, while the latter focuses on building connections between employees on a team, office, and company level. They are both comprised of employees from across the organization. I encourage you to read this blog from earlier in the year describing the experience of one member:
We are also building increasingly robust education opportunities. This year our Diversity & Inclusion Working Group has been building a series of learning opportunities around allyship. Launching in July, each month features topics under the umbrella of allyship such as general skills, LGBTQIA+ allyship, countering ageism, and disability allyship.
The People Operations team continues to build education and reminders to catch biases throughout our people process. One example is our Interview Unconscious Bias training, a session where we help interviewers understand what biases can be triggered and how to prevent them throughout the interview process. Our performance feedback training also included learnings on unconscious bias. This integration of bias concepts into broader trainings will continue to be a central part of our strategy to increase diversity, inclusion and equity at Clover.
This year’s data:
Over the last year and a half, we have taken a couple of different approaches to collecting demographic data. This year, we collected data through an anonymous CultureAmp survey with the same approach as our released 2017 data, which included additional, more specific race/ethnicity categories than the EEOC data with which many of us are familiar.
Last year we had 83% participation, and this year we had 73% participation. Why the difference? We believe there are two key reasons. First, this year’s survey was much longer. It included twenty-five questions measuring employee experience dimensions such as belonging and fairness, while last year’s only asked seven. Additionally, our survey ran a shorter period of time. There are always trade-offs when designing survey timelines and we continue to learn more about balancing those trade-offs in employee survey initiatives.
Because of this difference, we cannot be as confident as we would like that our data is representative of our current state, particularly when tracking differences from year-to-year. However, we have made a commitment to transparency to which we remain aligned. Below you will find our updated data of race/ethnicity, gender, and age. Please note that identities with fewer than 5 members are not included in the below data for privacy reasons.
Gender:
There are a couple of key things to point out around our gender representation. The largest is that our data is significantly skewed by some of our larger departments, including our Experience team, teams who provide support to our members and provider network, and our Clinical team, which includes Medical Assistants and Nurse Practitioners. Together, those teams are over 40% of our organization and both have larger populations of women, well over 50% women each. That is why our resulting actions will be very different for different company branches. In these branches, we are finding significant skewing showing we lack strong representation of men and non-binary candidates in the initial applicants and as a result we are focusing on diversifying our sourcing locations and analyzing the language we use in job descriptions. In other teams, there are similar strategies but different sourcing partners being considered due to what we see in our applicant pool. Our philosophy is that talent is distributed across all demographics, including gender, and diversifying sources to include more people will increase our quality of candidates.
In technology, we do see a higher than technology industry average representation of women, but a notably lower percentage than last we saw last year. To demonstrate how much sample size matter here, our survey results showed 34% of people in tech were women and when you add in those who did not complete the survey that number goes down to 26%; our HR database (HRIS) shows we remain at 39% women in technology roles. Please note the chart below represents survey numbers prior to adjustment for non-participants.
We can see that gender ratio is consistent year over year in our HRIS, and that the 2018 survey is the outlier. We want to follow up to understand whether there were any systematic effects that made women less likely to respond, but we are confident that we have maintained a roughly 60/40 gender ratio. It will be important to keep the survey outlier in mind when analyzing responses based on gender.
Even with that seeming lack of actual movement, we will always strive to move closer 45%+ representation for women. Why not 50%? Because we would also like to make sure we are including non-binary genders in our recruiting process and are not targeting our representation limited to a gender binary. In leadership roles, we continue to focus on making sure that we are watching the difference between representation at entry leadership levels and executive levels. We have begun this work by setting expectations with our executive sourcing partners to provide a diverse range of candidates from a diverse group of sources. This simple ask has resulted in more diverse pool of candidates.
Obviously we want to track our progress and setbacks from year-to-year, but our sample size makes it hard to be confident in some of the trends we see when comparing 2017 and 2018 data. When comparing the two data sets, the degree of “Not Reported” representation is greater than the changes we see from year-to-year for most demographic categories, making it impossible to know the exact change. That being said, we are investigating the potential changes as flags when you compare the data sets and continue to test and iterate ways to be inclusive in our hiring practices. Our recent partnership with Atipica, an analytics tool to examine the diversity of our talent pipeline, will allow us to best examine this data and be strategic in our attempts moving forward.
Race/ethnicity:
Similar to gender, we again see our overall numbers skewed by the same departments, Experience and Clinical teams, and are making branch-specific recommendations for how to diversify our initial pipeline sources. Comparing 2017 and 2018’s has again proven challenging, but we know that even if every employee categorized as “Not Reported” was in an underrepresented group, something highly unlikely, we would still have work to do. And that’s enough information for us to address.
Skewing heavily toward any one sourcing tool will only block diversity. That is why we are excited to be vetting many options and have already identified potential partners. In the remainder of 2018 and moving into 2019 we are looking forward to solidifying partnerships and learning how we can continue to bring in amazing employees as we grow.
Age:
The lack of age diversity in tech companies is one of the most silently accepted problems in diversity and inclusion and that needs to end. That is why we are again releasing age representation data. We also recently launched an internal training on Ageism and worked through examples of how it manifests in an employee’s experience as well as the recruiting process. This is a session we will repeat and integrate further into interview trainings as our first steps toward building awareness to help mitigate bias and increase representation for employees in under represented age brackets. Our numbers do not seem to have changed much from last year, but that in itself is a problem we are now beginning to address.
This year we are adding two new categories to our reporting: veteran status and disability. Neither are typically included in company diversity reporting, but we believe they are equally important pieces of Clover’s diversity and inclusion initiatives. No set of demographics is exhaustive and our Inclusion Survey measures many additional identity dimensions. We are starting by adding these two additional dimensions in our public reporting.
Veteran & Disability Status:
First the good: Our disability representation is strong. Not only that, but it has grown organically. This is something I, as a woman with a disability, have always looked for and yet never experienced at a previous company. As excited as it makes me to see this number, I also see it as a call to action to keep doing better and not just a strength on which to sit back. Before this survey was sent to employees, we had already begun assessing our practices around disability inclusion by mapping our practices to Disability:IN’s Disability Equality Index (DEI). As this work continues, we will hold ourselves accountable to making sure that we are supporting all employees in part through the benchmarking created in the DEI.
And now the bad: Our representation of veterans is low. Very low. We have started vetting several possible sourcing partners to make sure we are presenting employment opportunities to veterans in hopes to diversify our applicant pool and therefore our employee population. We will post updates on our public blog as those relationships solidify and develop.
Beyond representation:
I mentioned that there were twenty-five questions related to measuring employee experience. Within Clover, each executive has taken time to look through data for their branch and is currently finalizing specific action plans to continuously improve on various measures of belonging, voice, and communication. We look forward to sharing more about our efforts here along the way. We encourage you to read our recent blog on the Clover 5 for more about how we are using employee feedback data to continuously improve our culture.
Moving forward:
Our strategy continues to focus on integrating best practices into our existing procedures and providing safe learning opportunities on topics related to diversity, inclusion, and equity. For representation, we will always continue to evaluate our processes to look for new bias opportunities and examining how diversity changes (or doesn’t change) throughout each stage of the candidate process. We have already found that some of our sourcing tools connect us to a significantly homogenous candidate pool, while others provide us with a more diverse pool of qualified talent. We will be diversifying our job boards and placing significant emphasis on these new partnerships in the next two quarters, constantly and eagerly learning along the way.
We are excited to share our progress and roadblocks along this journey and remain committed to proactively communicating these learnings. We’re always trying to improve and would love to hear your thoughts and feedback. Please comment below or send me an email at jessie.wusthoff@cloverhealth.com.
Data was collected between 6/7/17 and 6/25/17 via an optional survey administered by CultureAmp. Open to the entire company, then 527 full-time employees, (73%) Clover employees participated. Employees who did not participate in the survey are bucketed into the “Not Reported” category.
Clover’s Technology Branch is made up of the following departments: Engineering, Product Management, Product Design, IT, Security, and Data Science.
For purposes of this data, Clover’s Leadership is defined as a combination of people managers and employees with a director, VP, or chief title.
This story is published in Noteworthy, where 10,000+ readers come every day to learn about the people & ideas shaping the products we love.
Follow our publication to see more product & design stories featured by the Journal team.
