{"id":45,"date":"2018-10-01T22:20:49","date_gmt":"2018-10-01T22:20:49","guid":{"rendered":"http:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/?p=45"},"modified":"2018-10-16T21:23:33","modified_gmt":"2018-10-16T21:23:33","slug":"hidden-figures-how-silicon-valley-keeps-diversity-data-secret","status":"publish","type":"post","link":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/hidden-figures-how-silicon-valley-keeps-diversity-data-secret\/","title":{"rendered":"Hidden figures: How Silicon Valley keeps diversity data secret"},"content":{"rendered":"

This story was produced by Reveal from The Center for Investigative Reporting<\/a>, a nonprofit news organization. Get their investigations emailed to you directly by signing up at revealnews.org\/newsletter<\/a>. Originally published<\/a> on\u00a0October 19, 2017.<\/p>\n

When the popular messaging platform Slack won a fastest-rising startup award last year, the company sent four black female engineers to accept it.<\/p>\n

Onstage at the TechCrunch awards show, one of the women praised Slack\u2019s diversity, citing<\/a>\u00a0a statistic from the company\u2019s\u00a02016 diversity report<\/a>: 9 percent of Slack\u2019s engineering team were black, Latina or Native American women.<\/p>\n

\u201cTHIS Is What Diversity In Tech Should Look Like,\u201d said one\u00a0HuffPost headline<\/a>.<\/p>\n

\"\"

At the 2016 TechCrunch awards in San Francisco, four black female engineers at Slack \u2013 from left to right, Megan Anctil, Erica Joy Baker, Kin\u00e9 Camara (at microphone) and Duretti Hirpa \u2013 accept the award for fastest-rising startup. In her speech, Camara cited a number from a Slack employee diversity survey that the company later acknowledged was flawed. Credit: TechCrunch<\/p><\/div>\n

It turns out that number came from an anonymous employee survey that Slack later acknowledged was flawed. While the company had said 6.9 percent of its technical team was black, for example, this year\u2019s\u00a0diversity report<\/a>\u00a0admitted the number should have been 4.3 percent. No mention was made of women of color this year.<\/p>\n

How many women of color do work at Slack? The answer is on a one-page form Slack and all companies with 100 or more employees send to the federal government each year. The forms \u2013 called\u00a0EEO-1 reports<\/a>\u00a0\u2013 show hard numbers of employees broken down by race, gender and job categories such as professionals, managers and executives. But Slack won\u2019t make it public.<\/p>\n

Even as California\u2019s Silicon Valley struggles with diversity and discrimination, most of the area\u2019s tech companies won\u2019t share that basic data with the public. Reveal from The Center for Investigative Reporting sought the government-mandated\u00a0EEO-1 reports<\/a>\u00a0from 211 of the biggest San Francisco Bay Area-based tech companies as part of an ongoing project examining diversity data in Silicon Valley.<\/p>\n

The requests included the\u00a0top 150 publicly traded tech companies,<\/a>\u00a0as compiled by The Mercury News in San Jose, and dozens of Bay Area \u201cunicorns,\u201d or private companies worth at least $1 billion, as estimated by research firms\u00a0Crunchbase<\/a>\u00a0and\u00a0CB Insights<\/a>. Of the 211 companies surveyed by Reveal, only 23 released their most recent reports. One of those 23, Clover Health, now says its report might be inaccurate.<\/p>\n

Still, the 23 reports represent the\u00a0largest public collection<\/a>\u00a0of EEO-1 figures that name Silicon Valley tech companies. A few private companies \u2013\u00a0Pinterest<\/a>,\u00a023andMe<\/a>,\u00a0View<\/a>\u00a0and\u00a0Clover Health<\/a>\u00a0\u2013 released their raw numbers for the first time. So did public companies\u00a0Square<\/a>, a payment processing platform, and\u00a0MobileIron<\/a>, which specializes in mobile security. Chipmaker\u00a0Nvidia<\/a>\u00a0also released its latest report exclusively to Reveal.<\/p>\n

Companies are under no legal obligation to release the reports, and the government keeps them confidential. But a recent push for transparency led some tech giants, including Google<\/a>\u00a0and\u00a0Facebook<\/a>, to share their raw numbers.<\/p>\n

Most others \u2013 including name brands such as Dropbox, Instacart, Netflix, PayPal, Pandora Media, Reddit and Tesla \u2013 still resist or put out basic pie charts that can be misleading, difficult to verify and impossible to compare. Three companies the federal government has accused of discriminatory hiring \u2013Oracle<\/a>,\u00a0Palantir Technologies<\/a>\u00a0and Splunk<\/a>\u00a0\u2013 also failed to disclose their demographics.<\/p>\n

For tech firms that did disclose, the numbers were particularly stark for executives.\u00a0Twitter<\/a>,\u00a0Square<\/a>\u00a0and\u00a023andMe<\/a>\u00a0did not report a single black, Latino or multiracial executive in 2016. Female executives who were black, Latina or multiracial were nonexistent at eight of the 23 companies, including Adobe Systems, Google and Lyft.<\/p>\n

The EEO-1 reports often are criticized for using clunky, outmoded categories. The multiracial option, for example, doesn\u2019t specify races. And the \u201cprofessional\u201d job category\u00a0includes<\/a>\u00a0both tech workers such as software engineers and non-tech employees such as lawyers, accountants and human resources specialists. Still, the raw numbers are the only standardized way to compare companies.<\/p>\n

Among professionals, Google and Apple had some of the lowest proportions of women, with 25 percent or less. Nvidia sat at the bottom, with 16 percent. Google fired<\/a>\u00a0an engineer in August over a memo arguing that biological differences make women less suited to tech or management. For all of the firms, the vast majority of female professionals \u2013 80 percent or more \u2013 were white and Asian.<\/p>\n

For black professionals, Hewlett Packard Enterprise had the highest proportion at 6.4 percent, compared with 1.7 percent at Apple and less than 1 percent at eBay. The online auction company also had the lowest proportion of Latino professionals.\u00a0Underrepresented minorities \u2014 including black, Latino, Native American, Pacific Islander and multiracial employees \u2014 made up 8.7 percent of Uber\u2019s professional workforce, compared to 14.2 percent for Lyft.<\/p>\n

Asian employees generally were much better represented among professionals than managers or executives, where the ranks were usually whiter.<\/p>\n

A tight hold on data<\/strong><\/p>\n

The battle over disclosing demographics is old news to veteran diversity advocates such as Erica Joy Baker. She argues that tech firms should share much more: promotions, departures, salaries and company ownership \u2013 all broken down by race and gender.<\/p>\n

\u201cWe\u2019re still having a fight about step one, we\u2019re still having a fight about the bare minimum,\u201d she said.<\/p>\n

Baker was onstage as a senior engineer at Slack when the company won the startup award last year. But she couldn\u2019t move up the ladder, she said.<\/p>\n

\u201cI started getting really demoralized and thinking maybe I shouldn\u2019t be a manager,\u201d she said.<\/p>\n

Baker got several outside offers and ended up as a senior engineering manager at the startup Patreon.<\/p>\n

\u201cSo, OK, it wasn\u2019t me,\u201d Baker said of hitting a ceiling at Slack. \u201cWho else is involved in that equation?\u201d<\/p>\n

Slack declined to say how many of its employees or managers are women of color. Its EEO-1 report would answer that question.<\/p>\n

\u201cAs we continue to understand and pursue best practices for advancing inclusive environments and reporting diversity data, we may consider publishing our EEO-1 reports in the future, but we are not planning to do so at this point,\u201d a Slack spokeswoman wrote in an email to Reveal.<\/p>\n

Fast-growing startups rarely have good processes for fairly evaluating and promoting engineers, which can hit people of color hardest, said Leslie Miley, a former engineering director at Slack and former engineering manager at Twitter. It\u2019s important, though, for companies to share their raw numbers, he said.<\/p>\n

\u201cHow can you get better unless you are truthful about what the situation is?\u201d he said. \u201cI don\u2019t understand why companies don\u2019t share those publicly.\u201d<\/p>\n

Diversity numbers rarely generate positive headlines, but they can make companies confront reality.<\/p>\n

\u201cInternally, it\u2019s that, \u2018Oh man, our numbers are coming out again, how do we look?\u2019 \u201d said Judith Williams, who was head of diversity at Dropbox and a diversity manager at Google. \u201cIt forces a conversation both externally and internally.\u201d<\/p>\n

Relying on percentages<\/strong><\/p>\n

If they disclose any numbers at all, most companies offer limited pie charts on diversity webpages or in corporate diversity reports. They say the government reports don\u2019t reflect how they view their own workforce. The firms invariably use percentages instead of raw numbers.<\/p>\n

\u201cI think they don\u2019t want people checking their math,\u201d Williams said.<\/p>\n

Some companies won\u2019t disclose numbers because they\u2019re afraid they\u2019ll get sued, said Pat Gillette, who was a defense lawyer for decades and is now a San Francisco-based mediator and speaker on gender diversity.<\/p>\n

\u201cThey feel like if they put it out there, they open themselves up to potential class actions based on the numbers,\u201d she said, \u201cwhich probably means they have a problem.\u201d<\/p>\n

Corporate diversity reports, she said, are easier to control than the EEO-1 form.<\/p>\n

\u201cThey can massage their numbers by how they define things. Because that\u2019s our job, that\u2019s what we do,\u201d Gillette said. \u201cA lawyer can make pretty much anything look pretty good.\u201d<\/p>\n

The CEO of cybersecurity company Lookout said on the company\u2019s diversity webpage that its numbers were \u201con par with our peers in the tech industry<\/a>.\u201d But its stated percentage of black employees, 0.4 percent, is lower than all 23 companies that disclosed their raw numbers to Reveal. Lookout removed the \u201con par with our peers\u201d language after questions from Reveal.<\/p>\n

Lookout\u2019s diversity page said the company was \u201cmaking our diversity metrics public to keep ourselves accountable for change.\u201d But Lookout wouldn\u2019t provide the numbers it reports to the government, calling it sensitive information.<\/p>\n

San Jose-based Synaptics called the aggregated, nameless statistics \u201cpersonal and confidential information.\u201d The subject is so sensitive that one prominent startup wrote in an email: \u201cOff the record, I can tell you (the company) is proud of it\u2019s (sic) inclusivity.\u201d<\/p>\n

Oracle \u2013 which is facing\u00a0government allegations<\/a>\u00a0that it favored Asian job candidates and paid white male employees more than others with the same job title \u2013 provides a\u00a0basic gender breakdown<\/a>\u00a0of its workforce but ignores race. Palantir Technologies, which\u00a0paid nearly $1.7 million<\/a>\u00a0earlier this year to settle charges that it discriminated against Asian applicants, has a\u00a0diversity webpage<\/a>\u00a0without any numbers. Tesla, fighting suits over\u00a0sexual<\/a>\u00a0and\u00a0racial<\/a>\u00a0harassment, doesn\u2019t share basic demographics either.<\/p>\n

Instead of releasing raw numbers, Varian Medical Systems pointed to a\u00a0company report<\/a>\u00a0that included percentages for only its 2016 summer intern program, which it said was 42 percent female and \u201c54% ethnically diverse.\u201d<\/p>\n

San Francisco-based Sunrun referred to a solar industry\u00a0diversity report<\/a>\u00a0\u2013 paid for in part by Sunrun \u2013 that said the company was building a \u201cbelonging culture\u201d without any specific numbers.<\/p>\n

Even companies that pat themselves on the back for diversity didn\u2019t want to talk about numbers. Workday, an HR software company, stated that it \u201cviews diversity as a business imperative\u201d in\u00a0announcing<\/a>\u00a0a new director of belonging and diversity in 2016. But a spokeswoman declined to provide Workday\u2019s EEO-1 report.<\/p>\n

When asked for an interview about the company\u2019s diversity efforts, spokeswoman Allison Kubota said in an email: \u201cWe don\u2019t have the resources.\u201d<\/p>\n

Two years ago, Udacity, which provides online courses,\u00a0announced<\/a>\u00a0that it had commissioned a study to \u201cdetermine how we stack up when it comes to diversity matters.\u201d<\/p>\n

A\u00a0company blog post<\/a>\u00a0in December 2015 said, \u201cIn our next post in our diversity series, we\u2019ll share with you the results of this study.\u201d Then the company has been silent for nearly two years: no new diversity posts, no study results.<\/p>\n

Udacity spokeswoman Amy Lester said the company has been too busy: \u201cWe\u2019ve had a ton of growth and many product launches over the past few years and had planned to post something sooner.\u201d This month, Lester wrote by email that Udacity would publish an update \u201cin the next month or so.\u201d<\/p>\n

The company wouldn\u2019t release its demographic numbers. \u201cWe don\u2019t believe the past EEO reports reflect the current state of the company at this time,\u201d Lester wrote.<\/p>\n

A blind eye to women of color<\/strong><\/p>\n

Even when companies put out diversity reports, they almost always have a big blind spot:\u00a0women of color<\/a>. There\u2019s usually a pie chart for race and another for gender, but rarely anything mentioning the representation of nonwhite women. (GitHub is a\u00a0rare exception<\/a>.)<\/p>\n

\u201cWomen of color experience the most bias and most marginalization,\u201d said Erica Joy Baker, of Patreon. \u201cThat\u2019s a good metric for people to see and know, especially if you\u2019re using those data points to see where you\u2019re going to work.\u201d<\/p>\n

It\u2019s part of a larger problem with diversity programs, said Y-Vonne Hutchinson, founder of ReadySet, a diversity solutions firm in Oakland.<\/p>\n

\u201cBecause they are not thinking of women of color, they are not addressing the problems they face,\u201d she said. \u201cDisproportionately, the beneficiaries of these initiatives tend to be white women.\u201d<\/p>\n

The government-mandated reports do account for women of color. But the reports have their own pitfalls. The only gender options are male and female, excluding those with a nonbinary gender identity. There\u2019s no accounting for diversity of age, sexual orientation or disability. And social media companies employ different kinds of professionals from chipmakers, ride-hailing companies or health startups. Clover Health, for example, employs nurses and social workers in addition to software engineers.<\/p>\n

There\u2019s another fundamental hitch: The numbers include guesswork. For all employees who decline to fill out forms identifying their race and gender, the company has to guess by looking at them. Judith Williams, the former Dropbox and Google diversity manager, said she\u2019s occasionally been called on for input when someone can\u2019t decide what race to label an employee.<\/p>\n

\u201cIt can be, \u2018I have a thousand people I have to identify, I have to get it done by the end of this week, I am just going through as quickly as I can,\u2019 \u201d Williams said. \u201cThey may have a when-in-doubt-just-click-this-box kind of thing. The problem is you just don\u2019t know.\u201d<\/p>\n

Even people who self-identify might check the wrong box as a joke or protest, though that happens rarely, Williams said.<\/p>\n

As with any data, there may be errors. Clover Health reported to the government that more than half of its executives were black women. It turns out that 31 black female administrative support workers were\u00a0accidentally\u00a0classified as executives, according to a spokeswoman. Also, 10 professional male employees classified as Pacific Islander should have been labeled Asian.<\/p>\n

Yet the official government numbers are the only reliable way to compare demographics across companies and years. Journalists have been\u00a0asking tech companies<\/a>\u00a0for their numbers for years, but were mostly\u00a0rebuffed<\/a>. (Newsrooms have their\u00a0own problems<\/a>\u00a0with\u00a0diversity<\/a>.)<\/p>\n

Then, in 2013, a Pinterest software engineer named Tracy Chou\u00a0called for hard numbers<\/a>\u00a0on female engineers and started a spreadsheet to track it, causing some companies to open up. The next year, the Rev. Jesse Jackson Sr. started showing up at Silicon Valley\u00a0shareholder meetings<\/a>\u00a0and\u00a0writing letters<\/a>\u00a0to major tech companies, demanding that they release their EEO-1 reports.<\/p>\n

\u201cMy leader saw that letter and got nervous, because we know what Jesse can do when he gets organizing,\u201d said Rachel Williams, Yelp\u2019s head of diversity and inclusion. \u201cHe can embarrass an organization.\u201d<\/p>\n

Yelp released its\u00a0EEO-1 report<\/a>\u00a0in 2014, but hasn\u2019t done it since then. In a phone interview, Williams indicated that she would provide a more recent report, then failed to do so despite repeated follow-up requests.<\/p>\n

She criticized the government-mandated numbers as out of date and \u201cprobably not very interesting.\u201d She said the government forms were invented by \u201cmostly white men\u201d and can \u201cperpetuate a system.\u201d More important, she said, is whether people of color are \u201cfeeling as though they\u2019re included, that they belong.\u201d<\/p>\n

\u201cWe\u2019re missing a lot when we make people choose a particular box that\u2019s designated by the government,\u201d she said. \u201cIf you talk to real people who are part black, part Hispanic \u2013 whatever \u2018Hispanic\u2019 actually means \u2013 part Irish, what box do you think you\u2019re supposed to choose?\u201d<\/p>\n

Yet when Yelp put out a\u00a0diversity report<\/a>\u00a0this month, the company used the same government categories \u2013 with a lot less detail. Like so many other Silicon Valley firms, Yelp didn\u2019t account for women of color at all.<\/p>\n

Jackson, in an interview with Reveal, called Rachel Williams\u2019 criticisms of the EEO-1 report \u201cnonsense.\u201d \u201cThat\u2019s a diversion,\u201d he said.<\/p>\n

Some companies, Jackson said, have slipped backward instead of making progress since his initial campaign. \u201cWe\u2019ll have to intensify our actions,\u201d he said. \u201cTBD.\u201d<\/p>\n

Reveal editor\u2019s note:<\/em><\/strong>\u00a0The Center for Investigative Reporting receives funding from Google News Lab. All editorial decisions<\/em>\u00a0are made independently<\/em><\/a>; donors receive no preferential coverage and do not influence the direction or findings of our reporting.<\/em><\/p>\n

Journalism organizations have their own problems with diversity. CIR doesn\u2019t file an EEO-1 report because we have fewer than 100 employees, but you can find our demographics\u00a0here<\/a>. They were compiled as part of\u00a0a survey<\/a>\u00a0by the American Society of News Editors on race and gender in the newsroom. We\u2019re also working to increase the ranks of investigative journalists of color through the\u00a0Reveal Investigative Fellows<\/a>\u00a0program.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

When the popular messaging platform Slack won a fastest-rising startup award last year, the company sent four black female engineers to accept it.<\/p>\n

Onstage at the TechCrunch awards show, one of the women praised Slack\u2019s diversity, citing<\/a>\u00a0a statistic from the company\u2019s 2016 diversity report<\/a>: 9 percent of Slack\u2019s engineering team were black, Latina or Native American women.<\/p>\n

\u201cTHIS Is What Diversity In Tech Should Look Like,\u201d said one HuffPost headline<\/a>.<\/p>\n[caption id=\"attachment_46\" align=\"alignnone\" width=\"1024\"]\"\" At the 2016 TechCrunch awards in San Francisco, four black female engineers at Slack \u2013 from left to right, Megan Anctil, Erica Joy Baker, Kin\u00e9 Camara (at microphone) and Duretti Hirpa \u2013 accept the award for fastest-rising startup. In her speech, Camara cited a number from a Slack employee diversity survey that the company later acknowledged was flawed. Credit: TechCrunch[\/caption]\n

It turns out that number came from an anonymous employee survey that Slack later acknowledged was flawed. While the company had said 6.9 percent of its technical team was black, for example, this year\u2019sdiversity report<\/a>\u00a0admitted the number should have been 4.3 percent. No mention was made of women of color this year.<\/p>\n

How many women of color do work at Slack? The answer is on a one-page form Slack and all companies with 100 or more employees send to the federal government each year. The forms \u2013 calledEEO-1 reports<\/a>\u00a0\u2013 show hard numbers of employees broken down by race, gender and job categories such as professionals, managers and executives. But Slack won\u2019t make it public.<\/p>\n","protected":false},"author":27,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-45","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"acf":[],"_links":{"self":[{"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/posts\/45","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/users\/27"}],"replies":[{"embeddable":true,"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/comments?post=45"}],"version-history":[{"count":0,"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/posts\/45\/revisions"}],"wp:attachment":[{"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/media?parent=45"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/categories?post=45"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/publicknowledge.sfmoma.org\/civic-data-solidarity\/wp-json\/wp\/v2\/tags?post=45"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}