Startup Include Report on Our First Cohort
This report shares the aggregated results of our first Startup Include cohort. We share trends in demographics, employee satisfaction, and workplace experiences across the Startup Include companies over the course of eight months of the program. As we hypothesized, the startups were more diverse than your average large tech company, and over the past several months, their commitment and initiatives resulted in improvements in employee satisfaction and changes in hiring. CEOs who participated saw promising results and want to extend the program to continue discussions with cohort CEO’s on challenges and solutions, and to meet with future cohorts to share experiences and learnings. We know we are far from finished on our journey and will continue to push forward together.
We are encouraged by increased employee satisfaction levels showing improved perceptions of inclusion and inspired to forge ahead with this program by offering it to other CEO’s and companies. We want to better understand fairness concerns and clique exclusion, and to provide more specific recommendations around hiring diverse teams, and we will focus on those areas in upcoming cohorts.
Some positive highlights are in satisfaction levels and improvements, and in employee demographics:
- Employees at companies participating in Startup Include report high satisfaction overall and with diversity and inclusion efforts — and their satisfaction grew in most areas over time. Underrepresented people of color demonstrated a significant increase over time in their satisfaction with communication and growth opportunities.
- There was significantly less underrepresentation for two categories compared to the overall tech industry: Women were 46% of the sample versus 34% in tech, and transgender and nonbinary employees were 1% versus 0%, and Latinx employees were 12% versus 6%. (We did not include multiracial employees (11% v. 5%) in the last comparison but intend to in future reports.)
Some results were negative: Women, LGBTQA employees, and women of color had significantly more negative workplace experiences than their peers. We did not see significant growth in diversity of the employee base across several underrepresented groups. We plan to analyze these findings for insights on solutions.
We launched Startup Include last year to define and encourage best practices in metrics, data collection, and reporting across tech companies. The Startup Include inaugural program began with a cohort of startups working together over nearly a year to collect baseline and follow-up data, review cohort data reports, and form a peer network for CEO’s to discuss challenges and recommendations for improving diversity and inclusion at their startups. The companies, Asana, Clef, Managed by Q, Patreon, Periscope Data, Genius Plaza, Puppet, Truss, Twilio, and Upserve, ranged in size from 10 to 900 employees, with the average size being 200 employees. About half are based in Silicon Valley.
Seven startups participated in the baseline data collection in September 2016, with employees taking a survey measuring demographics, satisfaction, and negative experiences. Companies filled out profiles, including inclusion practices and policies.
Each startup received a baseline report describing employee demographics and satisfaction levels. CEOs and their key diversity and inclusion team members met as a cohort to share ideas and experiences. Project Include took a pause to build its own survey tool and database, which extended the time period between surveys from the planned six months to eight months.
We ran a follow-up survey across five companies in June 2017. A total of 719 respondents participated in the baseline survey and 344 respondents participated in the follow-up. We looked for significant changes in the cohort over eight months by comparing the companies who participated in both baseline and follow-up data collection. We met four times as a group for open discussions. Project Include also visited most startups to lead a company Q&A on diversity and inclusion.
In the follow-up survey, 58% of the respondents were white and 15% were Asian. For underrepresented groups, Latinx employees comprised 12%, while Black employees comprised 3% (Figure 1). In the follow-up survey, 53% of the sample was male, 46% female, and 1% transgender or non-binary (Figure 2). Less than 0.5% of respondents said they were Native American, Alaskan Native, or First Nations. Thirty-six percent of the sample are immigrants, 26% are veterans, 14% are parents, 11% identify as LGBTQA, 10% do not have a college degree, and 6% have a self-reported disability (Figure 3).
At follow-up, Asian and Middle Eastern employees were most likely to be in engineering roles, while Black, Latinx, and white employees were less likely to be concentrated in engineering roles (Figures 4.1 and 4.2) Women, transgender men, and non-binary employees are significantly less likely than male employees to hold engineering roles or to be in senior, director, VP or C-level roles (one of the top three highest paying roles along with finance/operations/legal, and marketing) (Figures 4.3 and 4.4). There were no significant differences in the percentage of women of color compared to white women, LGBTQA compared to non-LGBTQA, or underrepresented POC compared to white employees in higher paid roles and positions (Appendix 1).
The Startup Include cohort has a higher percentage of women and Latinx professionals than the industry average. While there appear to be some differences in the race and gender demographics of the cohort from baseline to follow-up, because we did not collect identifiable information for each employee, our analyses indicated that there were no significant differences in employee demographics, but we cannot be certain that any minor changes in demographics are due to increased workforce diversity of the sample. It is promising, however, that among employees hired in the past 6 months (during the time that companies were participating in the Startup Include program), there were significant differences in the racial/ethnic backgrounds of the employees hired within the last six months compared to the team at baseline. There were significantly fewer white employees and more Asian employees. Increases in the percentage of women, nonbinary employees and transgender men, and in the percentage of Latinx employees were encouraging, though not statistically significant. No other significant racial/ethnic differences or gender demographic changes were found (Figure 5 and Figure 6).
Employee satisfaction over time
Six categories of employee satisfaction were examined at both baseline and follow-up: Overall Satisfaction in Job & Company; Fairness; Growth & Advancement; Communication & Decision-Making; Belonging; and Diversity & Inclusion. A summary of the mean for each category is provided in Figure 6; each category mean ranges from 1 to 5, with the higher numbers representing greater positive responses across the scale items (1=Strongly Disagree, and 5=Strongly Agree). Employee satisfaction was compared between baseline and follow-up with the following results:
- Employees were highly satisfied with their company’s diversity and inclusion efforts in both surveys; it was the most positively rated category.
- Employees’ perception of belonging was the second highest rated category at follow-up.
- Satisfaction with communication and perceived opportunities for growth and advancement increased significantly.
- Perception of fairness continued to be the lowest rated satisfaction scale, without any statistically significant change.
- At the company level, growth in satisfaction was as high as .27 points for Communication and .19 points for Diversity and Inclusion, demonstrating that some startups showed areas of significant growth beyond the cohort averages.
Comparing satisfaction levels over time revealed:
- Significant increases in perceptions of fairness in performance reviews, satisfaction with opportunities to increase professional skills, satisfaction with decision-making processes, and with work-life balance and the company’s effort to increase diversity and inclusion.
- Significant decreases in satisfaction with the company’s ability to motivate employees in their roles and the inclusion of their perspectives into decision-making processes.
Satisfaction by Employee Subgroup
We also analyzed: 1. differences in growth from baseline to follow-up by subgroups, and 2. differences by subgroups at follow-up (Appendix 2). We discovered:
- Underrepresented people of color demonstrated a significant increase in their satisfaction with communication and growth opportunities.
- Employees without four-year college degrees demonstrated a significant decrease in perceived fairness from baseline.
- Women, transgender men, and non-binary employees demonstrated a significant decrease in perceived fairness from baseline.
- Women, transgender men, and non-binary employees were significantly less satisfied overall, and had lower perceived fairness, satisfaction with growth and development opportunities, and satisfaction with communication and diversity and inclusion efforts than male employees.
- Employees with children were significantly more satisfied overall, satisfied with growth opportunities, and satisfied with diversity and inclusion efforts than employees without children.
- Veterans had significantly lower levels of belonging than non-veterans.
Employee Experiences Over Time
Data on employee experiences were collected across 6 categories: Unwanted sexual attention; Negative assumptions about ability; Exclusionary cliques/networks; Mistaken identity; Unwanted stereotypical comments; and Other. The percentage of employees reporting experiences are shown in Figure 7. We analyzed whether negative employee experiences changed from baseline to follow-up. Since baseline, there have been no significant changes in the percentage of employees who have experienced forms of unfair treatment, though we did note a high level of negative experiences with cliques, especially at follow-up.
Differences in Employee Experiences by Subgroups
We also looked for: 1. changes in negative experiences among demographic subgroups; and 2. significant differences in negative workplace experiences for subgroups at follow-up (Appendix 3).
Differences in quantity of experiences included:
- Significantly more underrepresented people of color experienced exclusionary cliques at follow-up (compared to baseline).
- Significantly more women of color experienced exclusionary cliques at follow-up (compared to baseline).
- Fewer employees with disabilities reported unwanted stereotypical comments at follow-up.
- There were no other significant changes in negative experiences among other subgroups.
- Women, transgender men, and non-binary employees experienced significantly more negative treatment overall, and more negative assumptions about skills and abilities than males.
- LGBTQA employees experienced significantly more negative workplace experiences overall than non-LGBTQA employees.
- Women of color experienced significantly less unwanted sexual attention than white women.
- Parents experienced exclusionary cliques significantly less than non-parents.
- Employees with disabilities experienced negative assumptions about skills and abilities and mistaken identity significantly less than their peers.
- There were no other significant differences in workplace experiences among other subgroups.
Over the past eight months, Startup Include company employees report high overall satisfaction and are highly satisfied with diversity and inclusion efforts. Satisfaction with communication and decision-making increased significantly, fairness decreased significantly, and employees’ satisfaction in fairness in performance reviews, decision-making processes, work-life balance, and the company’s effort to increase diversity increased significantly from baseline to follow-up.
There is still a lot of work to do, as the data reflect many of the patterns of underrepresentation in the tech industry. But the improvement over large companies’ demographics is encouraging, as are the many changes between the baseline and follow up reports, though it is too early to know for sure what caused all the changes. In the most recent survey, we saw the percentage of Latinx employees increase to 12% of the sample, while Black employees dropped to 3%, possibly due to an increase in “more than one” by 4%. In the follow-up survey, 46% of the sample were female, an increase of 5% from 8 months earlier. Ten percent of the follow-up sample includes employees who have earned less than a college degree, and 26% are veterans. Thirty-six percent of the sample are immigrants, 14% are parents, 6% have a self-reported disability, and 11% identifies as LGBTQA.
We also note that positive changes were accompanied by some negative data. Women, transgender men, and non-binary employees are significantly less likely than male employees to hold engineering roles or to be in senior, director, VP or C-level roles. Latinx employees and women were significantly more likely to hold low-paying roles than high-paying roles. Non-binary respondents decreased from 2% to 1% over the 8-month period.
Moving forward, we want to better understand the phenomenon of cliques and look for recommendations on how to lessen their exclusionary effects. We want to understand why perception of fairness in the follow-up survey did not change significantly even though there were increases in satisfaction in many other areas of fairness and in general. As we get more data points, we would like to look at intersectionality more closely. And we want to continue to improve recommendations for companies both on recruiting more diverse teams and inclusion across all groups.
We appreciate the time, continued commitment, and efforts of the CEOs, other execs, and employees who participated in the Startup Include program, and we look forward to working with these companies and others to collect more data that will help us improve recommendations and eventually design standards and benchmarks.
We also want to thank Dr. Allison Scott of the Kapor Center and Elizabeth deRenzy for gathering and analyzing the data and drafting this report, George Pang for singlehandedly writing a survey tool in four months to protect participant privacy, and Amanda Lenhart and Mary Madden of Data and Society and Dr. Freada Kapor Klein of the Kapor Center for reviewing the survey questionnaires.
And thank you for your interest in our results (and for reading to the end of the report). We are just getting started and already have a list of changes for the second cohort. Please look for progress and improvements as we continue to use data to drive diversity and inclusion in tech.
 Includes African, African American/Black, Hispanic/Latinx, and Native American/Alaskan Native/First Nations.
 The timeline is longer than the planned six month gap, because we built our own survey platform; we wanted to protect employee privacy by owning the survey technology in light of new post-election risks. Companies running surveys decreased, because one startup was acquired and another changed CEOs. One company had an IPO and only participated in the first meeting. One company dropped out before the first meeting. Other startups used different survey tools that we did not aggregate in this report.
 The difference in respondents at baseline and follow-up is due to 2 fewer companies participating in follow-up; quantitative analyses showed no significant differences in employee demographics from baseline to follow-up across eleven categories, including age, race/ethnicity, gender, LGBTQA, disability, immigrant status, parental and/or other care providing responsibilities, veteran status, and education.
 Given that identifiable information was not collected for individual employees and we did not have 100% participation, we cannot determine whether any changes in demographic makeup of the cohort are due to changes in employees (hiring more diverse employees) or changes in the sample (more diverse set of employees completing the survey).
 Industry average was calculated using available EEO-1 reports across 20 top-grossing tech companies (based on annual revenue), and including both hardware, software, and internet companies. The follow-up race/ethnicity percentages total over 100%, individuals could select more than one race.
 Demographic categories analyzed include: age, race/ethnicity, gender, LGBTQA, disability, immigrant status, parental and/or other care providing responsibilities, veteran status, and education.
* Differences in means between baseline and follow-up are statistically significant (p< 0.05).
** New variables were added to the follow-up survey. Increase in mean between baseline and follow-up is only statistically significant (p< 0.05) when the new variables are included the scale.
Subgroup categories from the follow-up surveys were analyzed to determine if any differences in job role and job level existed by subgroup. All significant differences in means between subgroups on each variable are identified with a (+) or (-), with (+) indicating significantly higher values and (-) indicating significantly lower values (p< 0.05).
Appendix 1. Differences in Role and Job Level
Appendix 2. Differences in Satisfaction
Appendix 3. Differences in Experiences