Skype at 10: How an Estonian startup transformed itself (and the world)

Chapter 4: ‘Are you smoking?’

A couple of years into the company, after Skype’s usage had exploded, Zennström was hearing daily from users. One of his favorite stories is about a college student who made a Skype audio call to his mother. They were talking a while when the mother paused and said, “Are you smoking cigarettes?” He was.

“The quality was so good she could hear that he was smoking,” Zennström says. “I hope he stopped smoking, by the way, but that was a funny story. I have heard a lot of these stories, and met a lot of people who have used it for such important things. To me, that’s the most rewarding thing ever about Skype. There’s something for everyone.”

Skype experienced a kind of hyper growth that very few companies experience, says Mark Gillett, Microsoft’s corporate vice president for Skype and Lync. Gillett came on board shortly after Skype’s sixth birthday to help the Silver Lake Partners evaluate where to invest in and improve the company’s technology.

Your browser does not support the video element.

“A million downloads in its first few months of life – it was incredible growth, and that was pretty tricky for a small team to handle,” he says. “They were being pulled in lots of different directions, which means in the early years Skype didn’t commit enough people and resources to moving the product forward as much as they would have wanted and liked.”

In its first 10 years the company “punched the reset button” a number of times, including making a number of major changes to its original peer-to-peer model running on users’ PCs. To be able to support millions of users, Skype turned to cloud technology and the use of dedicated “supernodes” hosted in data centers. Other major Skype engineering resets included the company’s decision to extend support beyond the Windows operating system, and changing the Skype platform to be able to work on mobile devices.

“Those are just a few examples. Skype has never been religious about anything,” says Simon Longbottom, senior director of product marketing. Nothing was sacred – not rules, not code, not process.

“I remember (former eBay-appointed CEO) Josh Silverman saying we should write all those things that we think are set in stone and be prepared to destroy them,” Longbottom says. “Nothing should be set in stone.”

Skype’s technological infrastructure was like an ever-expanding Lego masterpiece with blocks of all different origins, colors, sizes and shapes. Employees continuously built and built, adding bright new blocks while simultaneously repairing or removing the ones showing wear and tear from heavy daily use. And if something wasn’t quite working? The engineers would patch it, or re-architect it, or tear it down entirely and build something new from scratch.

In its heady startup days, Skype turned to its tech-savvy early adopters for significant help. Tamas Henning was a high school student in Romania when his brother, a college student, introduced him to Skype. The two, both tech geeks, started using it frequently. They loved it.

“And we found bugs, of course, because it was a really new product,” Henning says.

The brothers posted those bugs on Skype’s Internet forum, and eventually got involved in its beta testing program. Beta testers were (and are) invaluable to Skype, and early on it wasn’t unusual for Skype to send its most valuable testers laptops and headsets, or to fly them to locations around Europe for a get-together. Which is how Henning, still in high school, ended up flying to visit Skype’s Tallinn office in the fall of 2006.

“It was overpopulated, people were working on bean bags and on the stairs, but the atmosphere was really fun,” Henning says. “It was a real good bunch of people who were there because they believed in the product and whatever they did, just wanted to make the product better. Everybody was happy, energetic and trying to make sure this thing they delivered was almost perfect.”

Skype Milestones slideshow

Ten years ago, applying the idea of peer-to-peer technology to the idea of making calls via the Internet “was visionary – the best idea anybody could have,” Henning says.

Servers and clouds didn’t exist in the same way at that time, and small technology companies didn’t necessarily have the funds to buy and operate huge server farms to power their products.

Using peer-to-peer technology meant that the processing, storage and bandwidth needed to power Skype could be shared and distributed among all of its users, or peers. Each user contributed a share of computing resources to help Skype run. In return, that user could call a grandparent in Florida, a cousin in Antarctica, or a childhood friend in Japan and talk for hours.

The architecture that made Skype a success in the beginning became a problem to sustain as its popularity exploded. As time progressed, and Skype’s peer-to-peer network experienced multiple stomach-flipping outages, engineers knew something had to change.

“A simple mistake could make everything crumble,” Henning says. “People started to realize this was potentially not a sustainable way of maintaining a huge communication platform.”

There would have to be a shift in technology, which wasn’t easy considering all the factors. Skype would have to evolve its architecture in a way that was minimally disruptive to users, backwards compatible, friendly to multiple platforms and, eventually, be able to run on mobile devices as well.

“I know Skype had a lot of issues with its technology,” Henning says. “But even though we had these issues, the engineers working on the code are brilliant and the solutions they came up with and the whole architecture of Skype is actually quite amazing.”

Andrei Jefremov, the man hired to bring video to Skype, is one of those.

Jefremov had developed a video player for Kazaa, and was so impressed at working with Skype’s founders that he told them when they were ready to do video to give him a call. They did, and in August 2004, Jefremov started work to bring video calling to Skype. Though he was later joined by valuable developers like Karlheinz Wurm, for Jefremov’s first three years at Skype it was just him and a quality engineer. (By contrast, the video team now has upwards of 60 people.)

"There’s a chance for us to reinvent Skype today as long as we keep thinking about users, and that’s what we’re doing."
- Andrei Jefremov , principal architect at Skype

At the time, Jefremov was “the poster child for Skype.” He was living in London for work, and his wife was living back home in Stockholm. The two used voice calling to stay in touch, but Jefremov knew the experience could be better. In fact, he knew he could make it better – with video.

Less than a year after Jefremov started work, Zennström made the first public Skype video call in June 2005, and video calling launched publicly that December. To say it was a busy year for Jefremov would be putting it lightly.

He would later be joined by Karlheinz Wurm, now one of the highest ranking long-time Skype employees at Microsoft, but for a couple of years the video calling team was just Jefremov, a quality engineer, and a pool of enthusiastic beta testers.

If you saw him working during this period, you’d think he was stiffly dancing in front of his computers – leaning to the left, then to the right, left, right, left. He wasn’t dancing at all, but placing calls to himself between two computers and moving in front of each camera to make sure the video was working properly on both ends.

“People made a lot of jokes about the way I moved back and forth,” he says.

Developing video was a curious mix of working alone, with beta testers, and with engineers from around Europe – even the founders themselves. While Jefremov was bringing video to Skype calls, he was singularly focused on one question – what do users want? Years later, it’s still the key question.

In its first decade, Skype moved through a number of jarring changes not unlike a toddler becoming a teenager. And now, Skype is growing up.

“There’s a chance for us to reinvent Skype today as long as we keep thinking about users, and that’s what we’re doing. We’re putting users in the center,” says Jefremov, who is now the principal architect in Microsoft’s Real Time Media group (which includes Skype and Lync).

“A lot of friends ask the question of why are we writing code if Skype works so well on the PC,” he says. “The answer is: we are solving for today’s generation of problems and tomorrow’s, like making Skype and the mobile work together, and Skype and the cloud. This is where people are moving in their communication. It is extremely challenging, interesting and a lot of hard work. We look forward to solving it for our users.”

web statistics