When Facebook shelled out a stunning $19 billion for WhatsApp in February of last year, few here in the US had heard of the tiny Silicon Valley startup. The move surprised even the small coterie of journalists who so closely cover the Bay Area tech scene. This was because WhatsApp doesn't do much PR—and because its eponymous smartphone app was predominantly used overseas.
But, oh, how it was used overseas. Facebook shelled out $19 billion because the app—a way of sending text messages over the Internet, without paying the typically high SMS fees that wireless carriers charge to send texts across their private networks—had reached a similarly enormous 450 million people across Europe and the developing world, including India and Africa. Sure, the app was simple. But it met a real need. And it could serve as a platform for building all sorts of other simple services in places where wireless bandwidth is limited but people are hungry for the sort of instant communication we take for granted here in the US. One simple messaging app could provide voice calling, video calling, instant payments, and more.
"It's an amazingly powerful communication tool, distributed to populations that frankly have been reamed for years by SMS charges, and it has changed all that," David Soloff, who used WhatsApp to help bootstrap his company, Premise, told us shortly after Facebook acquisition was approved last fall. "This is a profound thing."
Now, a year after regulators approved the Facebook deal, WhatsApp is used by more than 900 million souls, which places it among the most popular apps on earth. Facebook is used by more than 1.5 billion, but few others even come close. And indeed, WhatsApp has added voice calling to its service, with more additions undoubtedly on the way. What may be most impressive, however, is that the company has done all of this with such a small staff. One of the most popular apps on the planet is run by a company that employs about 50 engineers.
And yet, the US media still has little to say about WhatsApp. The app is still mostly an overseas phenomenon—and the company still does little PR. In the one story they participated in—a profile in Forbes that appeared the day of the Facebook acquisition—company co-founder Jan Koum was unapologetic in saying he considers PR and press a drag on the company's time. "Marketing and press kicks up dust," said the Ukraine-born Koum, who founded the company alongside an old Yahoo colleague named Brian Acton. "It gets in your eye, and then you’re not focusing on the product."
But in the wake of the company's one-year anniversary with Facebook, Acton agreed to answer some questions, though only by email. The result is below, edited (slightly) for clarity.
In addition to what you'll read, we also asked Acton if the company plans to introduce video calling to Whatsapp. He didn't answer. But presumably, it will. Facebook already has added video calling to Facebook Messenger, the WhatsApp-like tool that Facebook built on its own. This is where the new breed of messaging app is headed.
Why does Facebook need two messaging apps? Well, they serve different parts of the world. Facebook Messenger—spun off from the primary Facebook app—boasts over 700 million users, and so many are here in the US. The two companies are a logical fit. Facebook can reach all those overseas WhatsApp users it can't reach with Messenger, and WhatsApp can plug into the vast technical infrastructure Facebook has built to serve its online empire. Facebook has not only built multiple computer data centers across the globe. It has built a network of its own machines inside many of the world's ISPs that can speed the delivery of information to the world's users, and it has purchased its own fiber cables for shuttling data across the planet.
WhatsApp takes good advantage of this global Facebook network. But in many ways, it remains its own company.
WIRED: WhatsApp serves over 900 million people, and most of them are overseas. The company built its operation in a way that's really the inverse of how most Silicon Valley tech startups approach the market. It seems like some of this happened "organically," but when and how and why did you go after the international market in earnest?
Brian Acton: In early 2010, we launched our first localized version of WhatsApp for iPhone. It included Spanish and German language translations, to name a couple. We felt from the very beginning that our product should be one that anyone could use no matter where they were in the world, and incorporating localized text was an obvious approach to that strategy. We found that every time we added a new language—or a new phone platform for that matter—we opened the doors to more users. Today, we support over 50 localized languages worldwide.
WIRED: You tackled the international market in part by inking deals with international wireless carriers, convincing them to bundle your app with phones as an alternative to classic SMS texting. How was this done? What was your pitch to the carriers? How easy—or how difficult—was this play?
Brian Acton: In general, we work with wireless carriers that understand that the mobile data services they provide are the future and who want to create more data usage—and customers!—within their network. WhatsApp is a great springboard app to that effect. Constructing our deals such that they are win/win/win—for our users, the carriers, and WhatsApp—has been the best way to pitch and win these relationships. Of course, each carrier is unique and speaking to each carrier individually is time-consuming. It makes the process more difficult than it should be.
WIRED: Are these deals the main reason the company has reached so many people? How else do you drive the app's expansion?
Brian Acton: We work hard for our app to work on as many phone platforms and carrier networks as possible. These deals have been a strong component toward our overall strategy, but so has focusing on app simplicity and performance as well as service reliability. All of these tactics interrelate and help foster growth across the world.
WIRED: Well, one thing's for sure: You've reached those 900 million with the help of a relatively small engineering staff. Today, the company still employs only about 50 engineers. What makes this possible?
Brian Acton: We hire the best talent that we can, and we stay focused on a core feature set of capabilities. We employ an engineer’s mindset and try to keep our service operating costs as low as possible (low server count, high quality hardware, minimal impact to staff productivity).
WIRED: You build the operation with the FreeBSD operating system and the Erlang programming language, two tools that aren't that commonly used in the Silicon Valley tech world. How and why did this happen?
Brian Acton: FreeBSD happened because both Jan and I have experience with FreeBSD from Yahoo!. FreeBSD has a nicely tuned network stack and extremely good reliability. We find managing FreeBSD installations to be quite straightforward. We came upon Erlang in a more indirect manner. Our original chat servers were built on Erlang, and we were able to leverage Erlang language features and evolve our service while at the same time maintaining very good uptime. Every step of the way, Erlang was rock solid and performant. I imagine that if we had come across significant hurdles or roadblocks along the way, we would have likely abandoned Erlang for a different language. Luckily, that never happened...
WIRED: Why is FreeBSD advantageous? Wouldn't Linux, an open source operating system that's far more widely used, be an easier option in some ways?
Brian Acton: Linux is a beast of complexity. FreeBSD has the advantage of being a single distribution with an extraordinarily good ports collection. To us, it has been an advantage as we have had very few problems that have occurred at the OS level. With Linux, you tend to have to wrangle more and you want to avoid that if you can.
WIRED: Why is Erlang so beneficial? Because it's designed for the type of communication you do? Because it handles concurrency so well?
Brian Acton: Erlang has shown great benefit to us. Neither Jan nor I had any exposure to Erlang before WhatsApp. Nevertheless, what we found is that language had solid foundations in industry, and it has served us well. It is true that Erlang is designed for near real-time communications. That said, Erlang is a generally good and useful general purpose language. There was serious thought and consideration that went into its construction. As one example, we've seen great benefits in high concurrency situations. We've also seen the ability to maintain great uptime as part of its hot code loading capabilities.
WIRED: Can other companies learn from the work you have done with FreeBSD and, in particular, Erlang? Are these part of the reason you can serve 900 million with only 50 engineers?
Brian Acton: Erlang and FreeBSD are great tools. However, as everyone knows—even if you give an ordinary chef the finest knives and cookery—he may not make the best meal. Frankly, it takes great and talented people to do what we do. I’ve had the very good fortune to work with the best in the industry, and I value each and every engineers’ contribution to what we have accomplished.
WIRED: You're now offering voice calling as well as instant messaging. How widely used is the service for voice?
Brian Acton: Through WhatsApp Calling, users can now make Internet VoIP calls, and this was a major initiative for 2015. We started the rollout back in January, and we continue to update and improve the service month over month. We continue to focus on service reliability and quality just as we have done for messaging over the past five years. The service is used widely in all parts of the world. We continue to see extremely good growth and usage.
WIRED: Are there changes you've had to make to your infrastructure as you've moved beyond messaging to voice? What changes?
Brian Acton: The biggest change was the addition of voice relay infrastructure. The nice part of that is that we were able to build and deploy this inside of Facebook’s global network and, as such, did not have to make significant changes to our own core infrastructure. Of course, building a voice product is not trivial and we had to make substantial changes to our mobile clients to support real time voice calling. All in a day’s work. :)
WIRED: You mostly operate independently from Facebook. You have your own office in Mountain View, California. But, indeed, you're starting to take advantage of Facebook's vast infrastructure. How else are you doing so?
Brian Acton: The best part of working with Facebook has been the cross-fertilization of ideas, people, and technology. With our voice launch, we were able to take advantage of Facebook’s worldwide network infrastructure. This was a big win for us as we had not yet made that investment in our own infrastructure. Continuing in that theme, Facebook has made significant investments in infrastructure that as a small company we were never able to make. We continue to learn and adopt Facebook technology every day.
WIRED: How can you take advantage of Facebook infrastructure in the future?
Brian Acton: There is the obvious choice of moving our hosting within Facebook’s infrastructure. This is a move that likely will take a good amount of time, as we would want to do it with zero customer interruption. In the near term, we focus on key wins where Facebook has infrastructure that we can leverage right away. Beyond that, Facebook has made big investments in large volume storage, data, and analytics, as well as consumer facing services (e.g., Places) that we seek to leverage going forward.
WIRED: Are there lessons other companies can learn from the way you've integrated with Facebook? Or is that a unique situation?
Brian Acton: I think every acquisition is unique and different. The best strategy is to listen to the founders and follow their lead. Mark [Zuckerberg] and Sheryl [Sandberg] have been extraordinarily good and doing exactly that. They trust us to integrate intelligently while maintaining the continuity of the business. They put very little external pressure on us and simply encourage us to grow. It’s been a terrific strategy so far.
WIRED: Apparently, the company almost never has meetings. Engineers say that the company operates completely differently than any other company they've ever been at. How specially is that approach/environment different? Where did that come from?
Brian Acton: I would love to say the company has no meetings. The truth is that the company has a few meetings. As a general rule, we try to keep meetings minimized in content and duration. We hope to create an environment in our office where people can spend the vast majority of their time writing code, fixing bugs, and building a better product. Much of this developed organically as part of the culture and company that we have built. It also grew from deliberate steps that we took to keep our work environment quiet and high functioning. Personally, I like our work environment a LOT but some people might find the quiet environment a little unnerving at first. :)
