ESPN CTO Aaron LaBerge is taking me on a tour of the company’s newest, state-of-the-art production center at its headquarters in Bristol, Connecticut, just over 100 miles from New York City. Dubbed Digital Center 2, the 194,000-square-foot facility, which opened in May 2014, still has a shiny-new feel to it, and the sets for shows such as SportsCenter and NFL Live, laden with giant displays and other gadgetry, have even more of a high-tech sheen in person than they do when you see them on an HDTV at home.
But I haven’t ventured to Bristol to gawk at the aspects of the new facility that were built to bedazzle sports fans. Not primarily, anyhow. It’s the behind-the-scenes technology that’s brought me here–and mostly the tech that ESPN uses to wrangle the vast amounts of video consumed by sports fans via TV, the web, and apps.
“The heart of this building is all IP,” LaBerge tells me as we survey the place. That’s “IP” as in “Internet protocol,” and he’s talking about the company’s years-long effort to treat digital video, on an array of platforms, as a core technological component rather than a repurposed variant of content meant first and foremost for delivery to cable TV subscribers.
With five studios, 16 editing rooms, and six production control centers, “Digital Center 2 is the most technologically advanced content facility in the world,” LaBerge says. “It’s beyond state of the art and is designed with enormous flexibility so that it won’t become obsolete.” Every day, ESPN produces more than 60 petabytes of data, and DC-2 is involved in most of it.
Signs of the company’s classic means of delivery are still highly visible at the Bristol campus–most noticeably, the platoon of mammoth satellite dishes sitting outside. What LaBerge and his team have been building is tougher to spot, since so much of it consists of software, services, miles of cabling, and rows of servers. But it’s there. And without it, ESPN would be unable to evolve as the world’s sports-consumption habits do.
“When you think about ESPN maybe five years ago, it was really a linear television company with digital add-ons,” he explains. “We are now literally a content company with multiple forms of distribution and consumption. We don’t care where the content goes.”
“We’re integrating teams that are left brain and right brain in many ways,” he adds. “Broadcast television has a certain way to think about solving problems. And guys who write Internet software have certain ways to think about problems. We erase the mediums and focus on the content and the fact that we’re going to have multiple mediums. Getting those people together has been, organizationally, a challenge for me. But it’s also been exciting.”
Not that the company is itching to ditch everything about its current way of doing business. For decades, TV in its conventional form has been very good to ESPN, which is used to pay-TV providers forking over the highest fees in the industry–about $6 per subscriber–for its programming. In recent years, however, as more consumers have opted for lower-end cable packages or cut the cord altogether, its subscriber base has taken a hit. Disney, which owns 80% of ESPN–Hearst controls the rest–depends so much on the sports network’s profits that any evidence it might be maxing out leaves investors skittish.
Disney and ESPN executives periodically make news by saying they don’t rule out the possibility of eventually selling ESPN in some form directly to consumers, as HBO does with HBO Now. Still, the company remains deeply committed to the idea of being part of a bundle of channels, whether delivered via cable, a satellite dish, or, in the case of Sling TV and PlayStation Vue, over the Internet.
But it’s long been obvious that ESPN isn’t only a channel in a bundle. (Exhibit A: People watched video clips on its digital platforms 465 million times in April.) It’s already the number one sports digital media brand, capturing 30% of eyeballs across its various services and apps. And wherever it chooses to go in the years to come, it doesn’t want technology to be an obstacle.
Years ago, tales of Jules Verne and Buck Rogers were made of dreams and wild bits of imagination. Today, modern technology has taken those dreams and that imagination and turned it into a reality that allows us to bring a television picture into your home via satellite.
–ESPN founder Bill Rasmussen in 1979, explaining to the channel’s earliest viewers how it could beam sports coverage across the country in a mere one-fifth of a second.
ESPN has been a fact of sports life for so long that it’s easy to forget that it was always a technology company, even in the age before PCs were pervasive and the consumer Internet arrived. In 1979, when the fledgling channel went on the air, cable TV was just becoming a mass medium. The mere fact that a startup could get anything at all on the air–let along on a 24/7 schedule–was an accomplishment. Sixteen years later, the company got on the web with the original version of ESPN.com, initially developed by a startup called Starwave, which employed LaBerge and was acquired by Disney in 1997. (LaBerge joined ESPN at that time, then left in 2007 to cofound his own tech startup. He returned in 2013 and was named CTO in 2015, succeeding Chuck Pagano, the company’s first CTO and a veteran of its 1979 launch.)
In 2006, ESPN even tried to become a wireless carrier, introducing an ESPN-branded cellular service with ESPN-branded phones. The venture fizzled out in less than a year, but the company says that the experience it gained building mobile software and services proved invaluable once the smartphone era got going in earnest shortly thereafter.
In fact, ESPN found it liked apps so much that it started cranking them out with abandon, until it had created more than 50 of them, more or less by accident. (There was even an ESPN pinball app.)
That era is over. “We’re kind of unusual in that because of the power of ESPN and its brand, we can get people to download an ESPN application,” says John Kosner, the company’s executive VP of digital and print media, and therefore the person ultimately responsible for crafting the strategy that the technology built by LaBerge’s team supports. The company recently decided that while it was important for it to be everywhere, that wasn’t the same thing as trying to do everything. So rather than scattering the power of its brand willy-nilly, it decided to concentrate it in a few high-profile apps that had huge audiences and the potential to grow even huger.
Henceforth, in fact, ESPN intends to focus on just three apps. There’s the main ESPN one. There’s WatchESPN, its video-centric app, which can authenticate you as a paying cable or satellite subscriber and then unlock content you’re entitled to watch. And there’s fantasy sports, which the company is planning to squeeze into one unified app by the end of the year.
As always, the sports world that ESPN covers runs on an odd blend of dependable, long-range scheduling and anything-can-happen unpredictability. “I can tell you–and there are people whose full-time function is to tell you–what’s going to happen each and every day,” says senior VP for digital management Ryan Spoon. “That’s a really enviable position. Because outside of an election year, if you’re the New York Times or Washington Post, you don’t necessarily know what’s showing up. We have that core base and we can build experiences around it.”
“There’s a beautiful cycle underneath,” he continues. “Fantasy football? It’s going to be great, and it needs to be great by August 1 this year. And March Madness is March 17. There are countless examples of that.”
But for all the ways in which ESPN can plan ahead, it must produce content in real time, and doing so for multiple platforms only makes the challenge more complex. “Our world is live,” says Spoon. “That’s the TV strategy. And that is the digital strategy. What we do for last night’s Kansas-Baylor game has to be done really well, instantaneously. And most of it is really perishable. You need to satisfy that live game window–both how you broadcast it and how it manifests itself on a little screen, wherever you are.”
Take, for instance, highlight clips. They’ve been a key ingredient of the ESPN recipe from the start. For decades, the company selected and prepped them at a pace designed to meet the needs of TV programs such as SportsCenter, where highlights originated as a way to recap sporting events that had in many cases concluded hours earlier.
Highlights matter at least as much in 2016 as ever, but in an era that runs at Twitter speed, people want to see them now. Today, the ESPN highlight machine is powered by a system LaBerge’s team created called FastClipper. It involves staffers whose job is to watch live sporting events at special stations. When something interesting happens, they can mark it on the fly, whereupon a clip is plucked out of the action and made immediately ready for distribution. (From a technological standpoint, that is: The company’s agreements with sports leagues don’t always permit it to share highlights of events while they’re still in progress.)
LaBerge stresses that FastClipper is only possible because ESPN rebuilt its platform from the ground up to treat video as data. “This happens only because of what we’ve unlocked with our infrastructure,” he says. “We didn’t say, ‘Hey, we need a real-time clipping system.’ We created an IP facility and started building products on top of it and realized we can do this now.”
ESPN is experimenting with ways to identify highlights without human intervention, but it’s not in a rush to cut its in-house experts out of the loop. “People come to ESPN for our authority and our style,” LaBerge says. “These people are picking [clips] with their own judgment. The key here is that they can do it instantly, as the game is happening. And then distribute it instantly.”
Pushing out highlights quickly makes sports fans happy, and happy sports fans have a more direct connection to ESPN’s business interests than you might guess. “What we’ve found is, when that happens, we see piracy of our content go way down, because most people just want a low-friction way to share something,” LaBerge explains. “We’re to the point where we’re 10 seconds faster than their ability to go back, rewind their DVR, and clip something.”
On ESPN’s to-do list: identifying, creating, and distributing highlights designed not to please everybody, but special audiences–such as NBA fans in Australia, who follow local heroes such as Melbourne-born Dante Exum. “There are no highlights on those guys,” LaBerge says. “We’re thinking we need to be able to cut highlights for players that we don’t care about here, but that could instantly create an Australian NBA highlight clip for all the Australian players, and include that in product for Australia. Which we’re going to do.”
For ESPN, being digital is only partly about the fundamental technical efforts that allow it to slice, dice, and distribute video for the Internet age. It’s also about building user interfaces. I’m reminded of how big that challenge is when LaBerge ushers me into a testing area dense with phones and tablets from an array of manufacturers–all of which his team needs on hand to verify that its software will perform adequately on the devices people use around the world.
Like everyone who writes apps, ESPN, has discovered that shortcuts are dangerous. “If you take an Apple experience and just port it into Android, it’s not going to go well, for a bunch of reasons,” says Spoon. Now, having winnowed its apps down to a handful, the company is working to make sure that each one feels like a native part of the platform it’s written for.
The apps face serious competition not so much from other sports apps as the company’s own ESPN.com. “The website has had such traffic for so long that gaining traffic for the native app means trying to convince people that it’s a better experience than something they’ve used forever,” says Kosner.
The company is also thinking about all the social networks where sports fans seek information and is tailoring its content for each rather than pushing out one generic version of ESPN. “We have people right now who are focused on taking the best story on the site and crafting different experiences for how it gets distributed on Twitter versus Facebook versus Instagram versus Snapchat,” Spoon says. “And I don’t believe you can be successful by doing it any other way.”
As platforms get more powerful, it’s very much in ESPN’s interest to keep up. For instance, the old Apple TV box could handle a four-Mbps stream at 30 frames per second; the new one does six-Mbps at 60 frames per second, allowing for higher-res, more fluid video. Even text overlays are more legible. “This is the direction we’re taking all of our over-the-top experiences,” LaBerge says.
ESPN’s international reach complicates matters further. It’s not just that the content the company serves up must be tailored to local tastes. (In India it highlights cricket, backgammon, and chess.) “A lot of our international growth is happening in countries that still have less than a megabit per second on wireless,” LaBerge says. “We’re having to look at our data payloads, look at our interfaces, and create different versions for each country.”
Meanwhile, pay TV in its most conventional form, as provided via cable or satellite, is hardly frozen in time. On the hardware side, 4K resolution–the next resolution bump beyond 1080p–is already here. Late last year, one study said 4K TVs would reach 23% market share in the U.S. in 2016; another forecasts they’ll be in nearly 50% of homes by 2020.
From a production and distribution standpoint, however, migrating to 4K still isn’t a no-brainer–especially for live events. ESPN, like the rest of the industry, is still figuring out how to make it happen, and the timetable it needs to happen on.
LaBerge is not an advocate of resolution for resolution’s sake. “4K is four times the amount of information, the cameras are different, the editing software is different, the load to deliver it is four times larger,” LaBerge says. “Most TVs before three months ago didn’t have the HDMI capability to actually play it. It’s very, very tricky. And for what? It’s very hard to distinguish. And so our thinking right now is that you’ll likely see us do some key events in 4K. You’ll also see us probably distribute up-converted versions of what we have, and they’ll look beautiful.”
To prove the point in ESPN’s TV lab, he shows me the same footage in 4K and upscaled 1080p and asks me which is which. I squint, think, squint some more, and eventually correctly identify the 4K version. Which might be meaningful if I wouldn’t have a 50% chance of getting it right just by flipping a coin.
Besides more pixels, TV is also getting more frames, an advancement that lends itself well to capturing action-packaged sports. But even bumping video from the standard 30 frames per second of yore to 60 fps isn’t automatically a panacea. “Because there’s camera motion, because the objects are moving, there’s motion blur over that 1/60th of a second,” ESPN principal engineer Sam Reisner tells me in the lab. “With higher resolution, even at 60 frames per second, you’re just seeing higher-resolution motion blur. Which isn’t more detail.”
The other way in which conventional pay TV is evolving is a reaction to the slick, interactive world of phones, tablets, and the web. Providers are building set-top boxes that aim to offer something far more appealing than the ugly channel grids and button-festooned remote controls of the past. In multiple cases, ESPN is actively involved in determining what its own service should look like.
“Consumers are expecting app-like experiences,” LaBerge says. “They want to just go into their environment and watch content with as little friction as possible. It’s easy when you have an open platform like iOS or Android or the web. Well, now our [TV] partners are saying, ‘Hey, do something custom for us.'”
Set-top boxes are often based on aging technology, so it isn’t easy to build something that feels as snappy and modern as what you can get on the latest smartphones, but ESPN is learning to maximize their potential. “We’re trying to create these rich environments,” LaBerge says. “And we’re doing a good job of it.”
Speaking of the future of technology and rich environments, I asked LaBerge one question that, in mid-2016, seemed pretty much obligatory: Does ESPN want to create experiences for VR headsets such as Facebook’s Oculus Rift?
He made a face just thinking about it. “The thing with virtual reality is, say, you’re watching a sporting event. You put a virtual camera next to Jack Nicklaus and you record the full field of view. Put your glasses on”–he mimics strapping on an Oculus-tyle headset–“and if you want to watch the game, you need to do this for two hours. You start sweating, the thing fogs up. For sports, it’s weird.”
Here’s the thing, though: Thanks to the Rift, Samsung’s Gear VR, HTC’s Vive, and Sony’s PlayStation VR, virtual reality may seem like it’s synonymous with headsets. But VR and its close cousin augmented reality are far broader concepts. And ESPN has been leveraging them since 1998, the year that Oculus founder Palmer Luckey turned six.
That was when it began enhancing its NFL broadcasts by painting a yellow first-down line on the field, using a technology created by a Silicon Valley startup called Sportvision. The game was real, the line was virtual, and they melded into an high-tech experience that helped people get more out of sports without requiring them to don a wacky headset. ESPN is actively researching further augmentation of its broadcasts with digital elements, using modern technologies and concepts that, for now, are still secret lab projects.
For all the effort ESPN is putting into thinking about its future, part of its strategy involves acknowledging that it doesn’t know what’s next. “The day before the iPhone was announced, the Nokia 6600 bar phone was the number one selling phone in the U.S.,” LaBerge says by way of historical precedent as we end my day at ESPN in his office. “It had a 40-by-30 pixel inch screen, black and white, no Internet. Things are changing everywhere. It really is unknown. We are aggressively building products and technologies that are agnostic.”
“At the end of the day, you’re taking data and you’re formatting it. It’s fun. There are a lot of hard problems to solve.”