Using data to improve diversity — and business performance

Studies show that diversity improves performance of human teams. Not sure about beach huts, though.

Too often diversity discussions in business are framed as a zero-sum game: affirmative action versus meritocracy, minority versus majority, them versus us.

There are some hopeful signs that the tech industry is starting to realize that this is not the case. Google, Facebook, Twitter, Apple, and Amazon — among others — have all made a point of releasing their diversity numbers, at least insofar as diversity means “gender and ethnicity,” and have done so for two years in a row now, so we can see how little things are improving. At least they recognize it’s a problem.

More significantly, these companies are releasing this data without apology, and with a frank recognition that diversity is a goal worth striving for. It makes companies smarter, it makes them more sensitive to the needs of a diverse customer base, and it’s the right thing to do.

But, as anyone who writes about the topic will discover in the comments on social media about their work, there’s still a sizable contingent of people who believe that companies need to lower their standards in order to increase the diversity of their work forces.

Not so, says Joelle Emerson, the founder of a relatively new agency called Paradigm that applies data-driven social-science techniques to the challenge of helping companies increase their diversity and manage more diverse workforces more effectively. Clients include Slack, Airbnb, Pinterest, and Udacity.

In fact, Emerson said, numerous studies show that diverse teams are more innovative and better at solving problems than teams where everyone shares the same background, race, or gender.

I spoke with Emerson at a discussion on diversity earlier this week at Draper University. Incidentally, Draper was a great venue for this chat. Every time I’ve visited Draper University I’ve been impressed by the diversity of the students (they are truly a global, multiethnic, mixed-gender group) as well as their infectious enthusiasm, curiosity, and seriousness of purpose. They ask great questions, too, and they are unfailingly welcoming and polite, which is something you don’t always encounter in Silicon Valley. Say what you will about Draper’s goofy “hero” iconography, they are doing something right in this department.

So if diverse teams produce better results, why not just focus on results, and let the diverse teams shine through their own merits?

Actually, Emerson told me, at least one study has shown that the more meritocratic people try to be, the less meritocratic their hiring and promotion decisions actually are. In other words, people are more likely to give big raises to men and small raises to women if they’re told to base their decisions exclusively on meritocratic principles. It’s a phenomenon known as the paradox of meritocracy.

You can see that dynamic at work in Silicon Valley, where investors pride themselves on their “pattern matching” and “data driven” decision making, but still somehow overwhelmingly prefer to invest in founders that look like them. When VCs are 91.8 percent male and 77.5 percent white, that’s a problem.

So if companies want to be truly meritocratic, they need to take steps to make more objective hiring and promotion decisions. That should result in better business performance — and more diversity at the same time, since it will eliminate built-in biases.

One such technique is the blind audition, which I’ve written about before. In symphony orchestras where people audition for jobs from behind concealing screens, hiring managers are forced to pay attention to the only thing that really matters: how well the person plays their instrument. Similarly, blind auditions in a tech company can help managers focus on the work a person can actually do, such as writing or coding, rather than on their look or their self-presentation.

I’ve used blind auditions, with reasonably good results, in hiring journalists. I will say this, however: If you’re forced to focus only on the work, it does make the hiring process more laborious, because you actually have to read every work sample very carefully. But there’s no doubt that leads to fairer decisions.

Emerson herself has a handful of recommendations in a smart article on raising the bar in hiring. Her basic thesis: If you fix your hiring process, you’ll wind up with employees who are both more diverse and more talented. She recommends doing that by democratizing the job application process (for instance, by eliminating the advantages that certain groups have thanks to training on how to interview); focusing on job-related skills; and retuning your “culture fit” questions around aspects of culture that really matter, such as “would I enjoy working with this person?” rather than “would I hang out with this person after work hours?”

This being Silicon Valley, there are a number of startups aimed at helping tech companies with their diversity efforts (in addition to Emerson’s Paradigm). Gapjumpers helps companies conduct blind auditions for more objective recruiting. Textio uses AI and natural language analysis to improve the text of job listings, removing words that might discourage women or other diverse applicants. And Jopwell helps connect black, Latino, and Native American job candidates with companies that want to hire them.

The bottom line: Diversity is — and should be — good for business. Smart companies will embrace this approach and make themselves not only more inclusive, but higher functioning.

For more on these issues, follow VentureBeat’s collections of stories on diversity, gender, and race. And please let me know how your company is — or isn’t — tackling diversity in tech.


 

Originally published on VentureBeat

Using data to improve diversity — and business performance

I, for one, welcome our new surveillance robot overlords

A Knightscope K5 robot patrols an empty parking lot.


Knightscope founder and CEO William Santana Li is not modest in his ambitions.

He estimates that crime — of all kinds — has an annual global economic impact of a trillion dollars. That’s $1 trillion lost every year due to theft, vandalism, robbery, violence, and more.

He wants to cut that in half. And to do that, he’s building a fleet of surveillance robots.

The Knightscope K5 is a tall (about 5.5 feet), dome-headed, wheeled robot loaded with cameras and sensors. It rolls around at a slow speed, just a few miles per hour, and as it goes it emits a sort of drone-like humming sound. The sound is deliberate, so the robot doesn’t surprise people by rolling up silently behind them, but it also gives it a somewhat eerie presence — a sensation that’s probably appropriate, given that it’s watching your every move with a sensor suite that includes light detection and ranging (LIDAR) devices; high-definition, low-light video cameras; a camera designed to read and recognize the digits on license plates; directional microphones; proximity sensors; an inertial measurement unit; and a GPS unit.

The K5 looks a bit like a cross between R2-D2 and a Dalek, and that’s right where Knightscope wants it. The design challenge, Santana Li told me, is to make something that commands respect but isn’t frightening; that’s serious-looking but approachable. So, for instance, even though the technology might allow it, Knightscope is not going to make black-painted robots that zoom around at 20 miles per hour, because that would terrify people. Like police officers and security guards, the K5 aims to be a constant, calm, reassuring presence.

Knightscope seems to be succeeding in that. I find the robots eerie, but I’m also a born cynic. Most people apparently find them charming, and many people take selfies or family photos with the robots.

Now, Santana Li’s approach might strike some people as a little overly optimistic. Indeed, he had a difficult time convincing any Silicon Valley venture capitalists to back him. “Hardware is too hard,” is the refrain that many investors will tell you: It’s simply too expensive and too difficult to build a sustainable hardware business, particularly in a world where fast, professional, and cost-effective Chinese manufacturing is so dominant.

Santana Li has little patience with that. In a recent onstage interview I had with him at GSV Labs’ Pioneer Summit, he expressed exasperation at the crop of tech entrepreneurs who are tackling social media sharing and apps instead of challenging hardware innovations.

“It’s un-American for an entrepreneur to take the path of least resistance!” Santana Li yelled.

Undeterred by the lack of venture backing, Santana Li raised a total of $7 million from a State Farm-backed incubator, Flextronics, and NTT DoCoMo. The company said it has just four customers so far, but claims to have over 100 on its waiting list.

Santana Li is also impatient with the arguments that privacy advocates bring up. You don’t like having a robot rolling around your neighborhood, recording video 24-7? Well, think how you’d feel if you were a victim of crime, Santana Li said. Now what’s more inconvenient?

I don’t find that argument fully persuasive, but the analogy he draws with beat cops and security guards is a good one: Sometimes the mere presence of authority is enough to deter crime, and — given the recent push to put body cameras on police officers — it’s maybe not such a big step to having a fully autonomous camera-equipped robot rolling around.

Besides, for the most part, Knightscope is punting on the privacy and data storage issues: It will let its customers sort those questions out.

The pricing model is what’s perhaps most intriguing. Knightscope plans to rent its robots out for $6.50 an hour — far lower than the $20 per hour most security guards make. With that, you’d get 24-7 coverage, a web-based console, and a slew of features that can supplement, not replace, whatever security force you already have. Customers could include businesses (for monitoring a mall or a parking lot, for instance), neighborhoods, or maybe someday even police forces.

And if Santana Li’s ambitions pan out, it could turn into a very big business. According to the National Institutes of Health, the U.S. spends $179 billion a year on police protection, legal proceedings, and corrections. At $6.50 per hour, supplemental security monitoring might look like a very economical alternative to policing and security guards — and with billions already being spent, Knightscope has the potential to carve off quite a big slice of revenue.

Just as long as it can keep from freaking people out too much.

Originally published on VentureBeat

I, for one, welcome our new surveillance robot overlords

Netflix reveals the future of enterprise tech: Here’s why

This is the legacy: Data centers that most people don't need to think about any more.

I was sitting in a conference on enterprise infrastructure this week when I realized that the generational shift long promised by cloud advocates is finally, irreversibly underway.

That shift is away from “legacy” data centers built on x86 servers, VMware-managed hypervisors, SQL databases from Oracle, and storage hardware provided by EMC. Replacing all that are web-scale (or at least wannabe web-scale) technologies based on containers, commodity hardware, NoSQL databases of various kinds, and flash storage. The new infrastructure is cheaper, easier to scale up to large volumes of data and computation, and more flexible and agile.

But who really cares about that architecture, except the billion-dollar infrastructure companies that are about to take a giant hit in their valuations? And by that I mean Dell, HP, IBM, Cisco, Oracle, and, yes, EMC (which Dell is in the process of trying to buy). These companies might not quite be the walking dead, as Wired called them this week, but they are certainly headed for a world of hurt, which is why Dell is trying to buy EMC: It needs to shore up its legacy business.

Who cares about them, except their shareholders? Because it’s now possible to build a billion-dollar company without ever setting foot in a data center. You don’t have to care whether the datacenter is using HP and Dell hardware or some cheap commodity CPUs built to spec by the cheapest possible manufacturer. All you need are virtual servers you can spin up on a moment’s notice, the ability to deploy containerized apps into that environment, and support for the unstructured databases you need to handle the massive influx of bits you’re about to start collecting and will need to analyze.

Netflix shows what that looks like, and why — for now — Amazon owns such a big piece of that future.

Neil Hunt, the chief product officer and vice president of engineering for Netflix, was speaking at the Engineering Summit on Infrastructure, which had been organized by Engineering Capital, a small, enterprise-focused VC fund. Hunt talked about Netflix’s longstanding use of Amazon Web Services, the market and technology leader in cloud services. But it’s not just Netflix, Hunt said: Everyone is moving toward AWS.

“AWS is now the basic layer of compute services,” said Hunt.

Netflix is not just heavily reliant on AWS — it’s about to become completely dependent. Hunt plans to power down his company’s last data center this year, at which point Netflix will be running almost entirely on outsourced cloud infrastructures, mostly operated by Amazon. (It’ll still run its own content delivery network — CDN.)

Note that this timeline is new. Netflix originally said it would shut down its last datacenter in 2014, and then again this past summer, but the future sometimes comes a little slower than expected. That’s one aspect of enterprise infrastructure that will probably never change.

Still, Hunt would be happy to be less dependent on a single vendor: “That’s a somewhat uncomfortable place: To be dependent on a partner who is also competing with you,” Hunt said, referring to the fact that Amazon also sells a streaming video service.

But up to now, Hunt hasn’t found a single provider that matches Amazon in terms of its capabilities and scope.

“AWS is years ahead of Azure and a year or two ahead of Google in terms of the features and levels of abstraction they offer,” Hunt said.

“It’s getting closer, but Amazon keeps raising the bar in terms of AWS features.”

Gleb Budman, the CEO of Backblaze, which also recently began providing storage services that compete with AWS, asked Hunt if he’d consider using other cloud providers, even piecemeal. For the most part, Hunt said, the answer was no. Apart from a few tests here and there (Netflix is backing up data to Google, for instance), the company is almost entirely based on AWS.

So is the battle for the next generation over? Hardly. AWS has an enormous head start, but there is still no standardization of cloud services — something that Hunt believes will be necessary.

Hunt looks forward to a day when there is more standardization among computing components — and, by extension, more competition for AWS.

“Let’s get it right. Let’s make a standard toolkit that software engineers use when building software, just like hardware engineers use when building a bridge,” Hunt said.

“Then we’ll see an incredible increase in productivity.”

originally published on VentureBeat

Netflix reveals the future of enterprise tech: Here’s why

Twitter Moments joins a long lineup of attempts to curate the news

Twitter Moments screenshot 1


Twitter made its long-awaited move into the news business this week with the launch of Twitter Moments, a new tab in Twitter’s mobile apps that let you see semi-curated summaries of the biggest news stories, as represented by things people are tweeting.

It makes sense, given that Twitter contains — among the 500 million things people tweet every day — an enormous amount of “news,” however you define that. But finding the news you’re interested in has historically been very difficult. You need to spend a lot of time creating lists or following people who actually have newsworthy things to say, and even then, their smartest tweets are often mixed up with a whole lot of stuff that may be interesting, and even funny, but which hardly qualifies as useful information.

Twitter, however, is a latecomer to the social news curation game. Lots of people have attempted to extract useful signals about the news from the huge mess of social data, with varying results. Let’s put Twitter Moments in context:

Techmeme: One of the earliest attempts to bring order to the news, Techmeme focuses on tech news. Tech journalists have a love-hate relationship with it, and can become obsessed with “getting on Techmeme” to the detriment of actually producing useful, well-written news. But by aggregating stories from a variety of sources and giving prominent links to the most useful and/or most-referenced stories, Techmeme actually is a handy way to scan the day’s top tech news.

Google News: Less focused on social signals than textual ones, Google News uses its analytic tools to group together related stories and highlight the biggest ones. Unlike Techmeme, it’s entirely driven by algorithms, and that means it often makes weird choices. I’ve heard that Google uses social sharing signals from Google+ to help determine which stories appear on Google News, but have never heard definitive confirmation of that — and now that Google+ is all but dead, it’s mostly moot. I find Google News an unsatisfying home page, but it is a good place to search for news once you’ve found it.

Flipboard: The closest thing to a magazine experience on mobile, Flipboard arguably presents the most readable, news-centric view of your social stream by letting you view stories that people in your Twitter or Facebook networks have shared. Unfortunately, it doesn’t do a lot of filtering or weighting of those stories (to make the most-shared ones more prominent, for instance).

Pulse: LinkedIn has been putting a lot of effort into curating news, and Pulse shows some of the fruits of that effort. Its most useful feature is the ability to notify you whenever one of your LinkedIn contacts is mentioned in the news. It also presents a list of stories based on what people in your network are sharing, which can be handy — but that feed is often dominated by the kind of self-promotional stuff that many people on LinkedIn can’t stop posting. More relevant are the daily news roundups from LinkedIn’s editors.

Nuzzel: This app has been getting a lot of press lately, first because Twitter investor Chris Sacca suggested that Twitter ought to buy it. It’s not a bad idea: Nuzzel actually makes Twitter useful for news by looking at the URLs that the people you follow are tweeting. If enough of them tweet the same URL, it puts that story in your news feed on Nuzzel; if even more people tweet it, Nuzzel will send a notification to your device. That’s handy if you have interesting people in your Twitter feed who tweet about news you’re interested in, but Nuzzel also offers some curated lists that can augment that, and may be expanding its curated feeds soon. I like the Nuzzel experience a lot, even if its algorithm is relatively basic — showing that you don’t necessarily need high-order artificial intelligence to extract the news from Twitter.

Twitter Moments: If you want a TV-like experience showing you some cool pictures and videos from the top news, sports, and entertainment topics, this is the place to go. I’ve been using it for less than a day, since it was first released, but my initial impression is that this is a good way for Twitter to highlight interesting things without asking me to do a lot of work to find those things. The big drawback is that it’s entirely self-contained: None of the tweets link out to stories on the Web, so if I want to see more than just headlines and pictures, I have to go somewhere else.

Upvoted: One more site worth mentioning just launched: Upvoted, a homepage that Reddit has put together out of the stories posted on that social network. One key feature of Upvoted: It’s just the stories, no comments or votes allowed. In other words, if you love Reddit’s obsession with nerd culture, kitten GIFs, space exploration, and geek love stories, but you hate its toxic mix of racism, sexism, and juvenile stupidity, Upvoted is the place for you.

What conclusions can you draw from this admittedly biased and ad-hoc survey of the landscape? First of all, Twitter Moments is way behind the rest of the pack in terms of social curation capabilities. It is hardly the “bold change” Twitter execs want you to believe it is: It’s kinda neat, but ultimately underwhelming.

Second, nobody is really using algorithms of any sophistication, with the possible exception of Google News: All of these sites and apps rely on the most basic stats, such as how many times a URL is shared, and most of them — including Twitter — add a significant layer of human curation. You’d think it would be pretty easy for Twitter — or someone else — to come up with a more effective solution than that.

Third, news consumers still have to put a fair amount of work in before they can get the news they want, consistently and readably. There is still a big opportunity for a company that can figure out how to curate a set of news, tailored to each reader’s interest, with speed and reliability.

Whether there’s a business model in doing that remains to be seen, however: The news has not exactly been a good place to find high rates of return on investment in the past few decades, and it’s getting even worse as online advertising approaches the end of the line. But that’s a topic for another day.

originally published on VentureBeat

Twitter Moments joins a long lineup of attempts to curate the news