A few days after social network site Facebook opened its platform to third-party sites, iLike, a popular music sharing site has scored 723,936 users on Facebook. It is by far the most popular application. We interviewed iLike chief executive, Ali Partovi. He compares Facebook’s platform to the web itself.
VB: Tell me about your experiences with Platform so far. You’ve been working on putting iLike on Facebook for several months now. Yet on the integration since Friday morning, there have been bugs and other issues on iLike’s end. What’s the status? Do you have enough servers now?
AP: So, first to give you the back-story on how we got involved: Over the past several months, we’ve pushed and pushed with Facebook asking for some sort of exclusive relationship.
They repeatedly said they won’t do an exclusive relationship, but would rather create a level playing field where we could compete with other third parties. We then gave up a bit, and we were actually a bit late to the game learning about the platform in detail. But when we finally did get access, our President, Hadi Partovi (my twin brother) took very little time to decide this was a huge strategic priority. That was a month ago.
We re-prioritized everything else, and started moving our people off other projects onto this. First two or three people, then a few more, and by the end it was a huge group of engineers pulling back-to-back all-nighters for a week-long sprint to the launch.
VB: What made iLike think that Facebook Platform would be a big deal? What stood out about it?
AP: Hadi has a strong background in the concept of platforms… at 24 he became the head of product management in the IE group at Microsoft, and was a key player in the browser wars. A month ago, even though the Facebook Platform wasn’t fully fleshed out, he saw just from the early beginnings of it that this could redefine web development.
What he said was, “in the history of computing, there was the personal computer, there was Windows, there was the web, and now the Facebook platform.” You can imagine that I and most our company was pretty skeptical. But he makes these calls so we followed him.
As to what stood out: it’s a combination of 3 things:
1) The technology itself — Facebook’s platform, like any platform, offers the developer building blocks to build apps faster than they could if they were starting from scratch, and to tap into a rich source of data & capabilities that would never otherwise be available.
2) The potential for viral spread — due to the way the Facebook news feed works, an app can spread across the community entirely by viral spread, as friends get notified when one person adopts it… this essentially bypasses the idea of trying to make your app “viral” as a standalone, because the Facebook is itself naturally viral.
3) The rhetoric from the Facebook management team, starting from the CEO himself, made it clear that they have a long-term commitment to a level playing field. For example, they absolutely refused to give us any special advantage, insisting that the market needs to see a level playing field… we offered them ownership in our company, money, etc — but they had no interest. Furthermore, they built and launched their own “video” app, but left it to “compete” on its own merits alongside other third-party apps rather than making it “pre-installed” for all Facebook users.
So, #1 and #2 made this something we had to jump on, and #3 made us comfortable with the long-term strategic implications.
VB: Some Facebook applications are still more viral than others… tell me why iLike is so popular?
AP: To be honest, we’re not completely sure! There is certainly some art and science to making an app viral, but also a healthy dose of black magic
But for one thing, iLike is quite different from almost all the other apps, in that it’s not simply a “widget” to add to your page, but an entire expansion of the Facebook feature set to add a rich music experience. For example, artist profiles — our app contains 500,000 and growing profiles for artists from major label acts to unsigned artists. It does include the ability to add songs to your pages, but also much more that you get on the app pages themselves — e.g. see upcoming concerts, who else in the community is going, etc.
VB: You’re able to incorporate information from Garageband, right? [Garageband.com is a companion site to iLike that features profiles of independent musicians.] From that, what sort of traffic are you seeing back to iLike.com and Garageband from the FB platform?
Thanks to Facebook, the moment a new user visits iLike [within Facebook], the very first page they see is automatically personalized to them already. Thanks to [the information on your profile at] Facebook, on your very first page view, we know your music tastes, your location, your list of friends, and their music tastes. So we can immediately tell you, “here’s one of your favorite artists with a concert near you, and these are your friends who might wanna go.”
We’re accomplishing the same thing on our own website (www.iLike.com), but you can imagine it’s much harder: first you have to tell us your music tastes, then invite all your friends, then they have to tell us their music tastes… Facebook solves the classic chicken-and-egg problem for a wide range of social applications like this because they already have all the chickens and eggs ready for you to use.
VB: How do you plan to monetize your Facebook user base, now that you have one?
AP: Rather than viewing our Facebook app as a way to get people to click through to our site, we view it as a self-contained site in itself, where we might even be able to build a bigger business than our own site. So, regarding monetization: We already make money from the links to buy music on iTunes or buy tickets on Ticketmaster. We also plan to place banner ads throughout the iLike app on Facebook.
The traffic on iLike.com was already quite substantial (well over a million registered users — I’m not going to disclose yet quite how many, but we’ll be making an announcement soon)… and to our astonishment the traffic on our 4-day-old Facebook app is already larger!
In terms of daily signups, iLike on Facebook trounces anything else we do… iLike on Facebook has been signing up roughly 200,000 new members a day — the only thing I can think of on the internet that’s growing faster than that is MySpace and maybe some of the free email/instant messaging services. And we haven’t even brought out the big guns to drive our growth (marketing to the existing iLike base, or paying for marketing on Facebook)
As for GarageBand, the number of MP3 streams and downloads has tripled almost overnight. It’s a very good time to be a GarageBand.com musician.
VB: As to the other question, about the Platform terms of service — based on what you’ve said already, you don’t see a risk of FB trying to develop its own music app that it favors, or charging iLike for its use of platform?
AP: It’s a mutually symbiotic relationship. This is just the very beginning of the race, and our only concern right now is to put distance between us and whoever #2 might be. But the longer term question is, what if Facebook themselves wanted to build their own? The answer is that we’ll be adding value in ways that will be hard for them to duplicate. Or rather, hard for them to duplicate simply by building some code.
There are two other hypothetical long-term worries:
1) Facebook starts taxing us. I think the way they are already taxing us is by keeping the prime ad real estate (top and left side), which they sell; maybe they’d consider increasing their pixels of ad space. but I highly doubt they’d actually demand revenue: they’re smart enough to know that the way to make money is to keep your slice of the pie small, and make the pie grow huge… rather than try to increase your slice of the pie and risk shrinking the pie.
2) Facebook technically can’t keep up with the demands of web app development… what if the FB Platform somehow breaks under its own weight, i.e. they can’t actually technically support all the apps? That is what has been happening a lot the last few days… that one is the only legit risk, but I think it’s a good horse to bet on!
VB: What do you plan to add that will let you keep that lead? Do you think that iLike/Garageband is already unique — hard for anyone to duplicate?
AP: For example, we’ll be collecting a lot more about people’s music tastes (right now we merely scan what they’ve typed into their profiles; soon we’ll hook up our iTunes applet that scans your entire music library to learn your tastes). We’re also letting people click, “I’m going,” or “I wanna go,” for any concert — this is building a giant database that makes our app more valuable, and anybody wanting to duplicate that would be starting from zero.
Imagine hypothetically, you’re wondering who to invite to the Arcade Fire concert. You can go to iLike, where millions of people are actively declaring which concerts they are going to, etc. Or you can go to the newly launched “official Facebook music” app that has just started and doesn’t have many users. Our service will actually be better than any challenger, because of the strength in numbers, because of the rich data we’re collecting.
VB: Have you had any communication with Facebook, especially their legal department, about iLike’s relationship with record labels? Or other licensing-type issues?
AP: iLike has never had a single copyright-related issue, we secure all the necessary permissions from the copyright holders in advance, whether that’s the labels or the teenage bands. What’s great is that Facebook has created a platform where you can be the winner without breaking the rules.
VB: So, if this is to be a long-term problem, it would assume near-infinite growth in the number of apps and in the number of people who use them?
AP: Well, we’ve already had some outages, mostly because we weren’t equipped to handle such instantaneous popularity, but in some cases due to issues on Facebook’s side. On day one, we added 10,000 users in the first ten hours. Then, mysteriously, we were shut down… we called Facebook and discovered that we had tripped an internal limit they had set for an app that receives 100,000 page views in a day. So they increased the limit to 1 million. We were shut down again 3 hours later because we had already tripped the 1 million limit.
Or, more seriously, today we experienced multiple very frustrating outages, which significantly curtailed our growth, and they were due to Facebook’s Platform having some issues… what’s worse is that the error page linked to our support email, which resulted in a ton of complaints to us.
I have sincere faith that if there’s anybody who can pull this off, it’s the Facebook guys — they are absolute professionals and unbelievably gifted engineers. Nobody had any idea how fast this would grow.
VB: What percentage of the problems you’ve experienced has been because of FB platform scaling issues as opposed to iLike scaling issues?
So far, 20% of iLike’s problems have been Facebook’s fault, and 80% our own fault (or that of third-parties on whom we rely for components of our service). We had dozens of servers ready to deploy, but we had no idea that this thing would eclipse everything else we’re doing quite so rapidly.
After a “deer-in-the-headlights” period, we decided to disable some of the features to reduce our load.
We also went on an emergency run to add servers… itself quite a story, because it’s not very easy to find 100+ industrial-strength servers in the middle of Memorial Day weekend! We spent Saturday loading a 24-ft truck from floor to ceiling with servers, and we spent Sunday “shucking” the servers (i.e. removing them from the boxes and styrofoam packaging), which itself takes many, many hours to unpack 100+ servers… all in preparation for the coming week. Facebook had told us that Memorial Day weekend is one of the slowest periods of the whole year for them. This is why we freaked out on Fri night when we saw that rather than slowing down towards the night as you’d expect, our traffic was still growing exponentially.
VB: What if, as some say, Facebook is the next Google. Google was search but now it’s building all sorts of apps (70+ at my last count). You’re saying FB won’t move in this direction? Tell me more.
AP: I think the more appropriate comparison is Microsoft. As we see it, Platform is to ordinary HTML what Windows was to DOS.
VB: What do you expect to be issues this week?
AP: Scaling to meet the demand. The best analogy I can make is that the spread of an app on Facebook is akin to the chain reaction that occurs in the core of a nuclear bomb… one atom splits, and sends particles in all directions, which cause neighboring atoms to split, and so on… In like manner, one person uses iLike, which notifies all their friends… In theory, there’s no reason why the growth would taper off at all — unless we run out of server capacity to handle it, or unless we begin reaching saturation of the entire Facebook community.
It feels a bit like Little Shop of Horrors… we keep adding servers to feed the beast and it keeps getting bigger and hungrier. Fortunately there are only a finite number of people in the Facebook community! Although I suspect Facebook’s own growth will accelerate dramatically, because people like us will contact their entire mailing list telling them to sign up for Facebook. The whole thing may seem insane, but remember, what’s at stake for us is the opportunity to become the music app for Facebook, which in turn could actually give us a shot at becoming the #1 music service in the world. Note that Facebook is the #1 photo service in the world, the #1 invite/event service in the world, etc.
A month ago, if I told anybody we could unseat MySpace in music, they’d say I was crazy. Today, it seems not only possible, but actually like it’s on track to happen unless we screw it up.
And that is an unbelievable turn of events, extremely eye-opening for me as much as anybody.
VB: Based on what you’ve experienced with platform so far, what are the main concepts that hackers/entrepreneurs and investors should understand.
As to wisdom with respect to Facebook, what I’d say is that anybody who is currently involved in building a consumer-facing website should be thinking about whether they should be building a Facebook app instead.
To me, the developers who don’t ask themselves that question are like the people building multimedia CD ROM software in 1996 who didn’t ask themselves if they should be building websites instead… i.e. companies that distributed maps of the country on CD Rom, or encyclopedias on CDROM, etc.
I think the Facebook platform is an epic, tectonic shift, a paradigm shift akin to Windows replacing DOS or the emergence of the Web itself.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.