Intel has a new report out today. It’s not about semiconductors. It’s about diversity: how Intel is doing when it comes to women and underrepresented minorities on its staff. The results are mixed — some strong and some, frankly, failures. Still the sheer amount of information is exceptional, and a direct challenge to other Silicon Valley giants who’ve chosen to hide their data.
Be Engineers About Diversity
Let’s start with some numbers.
Intel set a goal last year: Of all new hires, 40 percent have to be women or underrepresented minorities (black, Latino, Native American). The company had never hit that level in the past. So for Intel, it was an ambitious goal. And the company reports today: It managed to exceed it, hitting 43.1 percent.
Intel CEO Brian Krzanich shares some of his motivation: “I have two daughters. They’re both technically very bright. I want them to come into a workplace that’s a better place than the way the workplace is today.”
To do that, he says, Intel has to open up about how it’s doing inside. “There’s nothing here [that’s] top secret or should not be shared with the rest of the world, in my mind.”
Other tech giants don’t agree. Google, Facebook, Apple, Microsoft — companies that value metrics so much — have not publicly stated any measurable goals when it comes to diversity hiring, or the retention of employees. They haven’t disclosed the numbers of new hires or of exits from their companies, by gender and race.
Facebook, Google and Microsoft say their goals are not publicly available. An Apple spokesperson says the company has purposely decided to not set goals.
Intel is now an exception. Today’s report gets in the weeds. The company aims to increase its external diverse hiring rate to 45 percent this year, and it’s establishing a new target within this goal of a 14 percent hiring rate for underrepresented minorities.
There’s a sense of urgency. By 2020, Krzanich says, Intel must reach “full representation.”
By that he does not mean the company will look like America or its global consumer base. He means it’ll reflect the available talent pool. Intel has a long way to go — currently at 75 percent male and a combined 86 percent white and Asian.
These are not flattering numbers. Still Krzanich is upping the transparency, and he challenges his industry: If you’re serious about diversity, be engineers about it.
“When you’re engineers you need that data to understand the problem,” he says. “The data is just that. It’s data.”
Thesis: Pipeline Not The Biggest Issue
Intel’s made headway — without any magic pills. The company has red carpet events where female or black Intel executives meet with potential hires of their gender or race, and offer jobs (not just cocktails or photo ops). An Intel employee recounts that when a woman declined a job offer in his division, managers chose to not pick a white man but to look for another diversity hire instead. Bonuses are now tied to diversity goals.
Based on results so far, Krzanich has a bold claim: Intel’s ability to hire diversely is proof that the so-called “pipeline” problem — the idea that there just aren’t enough good candidates out there — is overhyped.
“If the pipeline was such a big problem, I would have come back as a failure there,” he says.
Intel is failing in another respect: retention. The company is losing black employees at a faster rate than other workers. The CEO says he doesn’t know why. Neil Green, a vice president at Intel, is part of a black leadership group at the company. He offers a reason: The casual conversations that lead to promotions and coaching don’t happen because the relationships just aren’t there.
“From the folks that I’ve talked to, I think African-Americans get frustrated that they’re not progressing faster,” Green says. “They aren’t necessarily meeting with and being sponsored by the senior executives.”
Freada Kapor Klein is an investor who funds diversity-focused startups like Jopwell, which connects job candidates who are underrepresented minorities to tech companies. Klein says culture is key.
Tech companies don’t just make new engineers pass a coding test. They have to pass a “culture fit” test. That’s where a huge amount of bias creeps in, she says, as existing teams only want a unicorn. “They are looking for the one-in-a-million person who comes from a different racial, ethnic, cultural, gender background, but in every other respect is identical to the white and Asian men who work there,” Klein says. “That’s not diversity.”
Klein also points to a post last fall on Medium by black female programmer Erica Joy Baker. Baker, a former Google engineer, describes how “colorless diversity” campaigns can have the effect of advancing white women while pushing people of color further back.
In Silicon Valley, Intel is known for taking on one of the hardest technological challenges of our time: packing twice as many transistors into a tiny computer chip every two years or so. It’s called Moore’s Law. When asked which is harder — upholding that law or meeting diversity goals — Krzanich says with diversity, every human approaches the data with bias. “That’s different than physics. When I go solve a physics problem, I can get it down to ‘this thickness isn’t right.’ Or ‘these voltages are wrong.’ And those are more measurable and simple.”
Another difference, he says, is that Intel has met Moore’s law time and again. Diversity is unfamiliar, which makes it a riskier challenge.