Skip to content


The Story of Google Friend Connect: Google Cambridge’s First Wholly Home-Grown Product

Maybe the old story, it posted in 11 March 2009 at xConomy.com, but I found that’s still interesting. The main idea of this is the difference what you should do between when you’re a startup and when you’re under a giant (a big company in the industry). “A small team was allowed to run with a product idea, much like a boot-strapped startup might do—but when the project showed promise, the corporation stepped in with marketing clout virtually impossible to match in the startup world. If done right, that kind of one-two (startup-giant) punch can make up for the overall bureaucratic sluggishness that often permeates big companies”

In May 2008, Google moved into colorful (hey, it’s Google) new offices in the heart of Kendall Square. Governor Deval Patrick played ping-pong at the grand opening with Google site director Steve Vinter. And since then, Google Cambridge has grown to some 200 people spread over four floors, and about evenly split between sales and engineering.

But while engineers at the local Googleplex are working on key infrastructure components such as BigTable and MapReduce and consumer-facing offerings like Google Books and Google Images, only one effort to date has been conceived, developed, and released entirely from here in Cambridge, MA: Google Friend Connect.

GFC is a social media tool that makes it extremely easy for website owners to add social features to their sites—without the need to learn programming. These features include gadgets or plug-ins that allow visitors to automatically import their personal profiles from Google, Yahoo, and other places without having to do it manually for each new site they want to join, as well as widgets for letting users rate and review things. Website owners simply select the features they want, fill out an extremely short (three-item) form, and the code is generated for them to cut and paste into their site. The idea is that such as easy tool will make sites more interactive and interesting, thereby helping site owners attract and retain their users. (We’ll get to Google’s motives later on).

[Editor's note: the second sentence of this paragraph was amended to reflect the correct number of monthly GFC websites and users.] And it seems to be working. Since the public beta launch last December, Google Friend Connect has grown to some 8 million websites a month spread across what are estimated to be hundreds of millions of users. Google isn’t exactly sure how big GFC has gotten. “It’s big,” though, is how tech lead manager Sami Shalabi sums things up. And there is apparently more news coming. Google says it is planning a big announcement in the near future about GFC.

Besides being the only Google product entirely born, bred, and weaned here in Cambridge, I found the GFC story interesting because it illustrates how entrepreneurship can flourish inside a big company (yes, Google is officially a big company). A small team was allowed to run with a product idea, much like a boot-strapped startup might do—but when the project showed promise, the corporation stepped in with marketing clout virtually impossible to match in the startup world. If done right, that kind of one-two (startup-giant) punch can make up for the overall bureaucratic sluggishness that often permeates big companies (and yes, Google has been known to exhibit some big company bureaucratic traits).
Google Friend Connect TeamThe story of Google Friend Connect begins, fittingly enough, with two friends. Shalabi and Mussie Shore met each other in 1998 at Iris Associates, an IBM subsidiary where they worked on a suite of Lotus products, most notably Lotus Quickplace and LotusNotes. I visited the pair on the sixth floor of Google Cambridge, where their 11-person team (counting them) occupies a funky, open workspace in a prime corner area. We then retreated to a “huddle room” (what non-Googlers call a conference room) named Magic Hat. It turns out that all the huddle rooms on the sixth floor are named after beers or breweries, which is cool with the GFC crew because three of the team are microbrewers. (Click on the team image to enlarge it and see a full caption.)

In 2006, Shalabi and Shore got the entrepreneurial bug and, together with Martin Fahey, co-founded Zingku, a social networking startup that made it easier to share things like photos, polls, and invitations via instant messages, e-mail, the Web, and (especially) cell phones. You know who purchased Zingku in fall 2007 (for an undisclosed sum). And it was just a few months after

coming into the Google fold that Shalabi and Shore hit on idea for what became Google Friend Connect.

“We were really dabbling, trying to figure out how we fit…and how to contribute to Google,” Shalabi says. Initially, he and Shore figured they would work on mobile products for Google. But they saw such work as incremental for both Google and themselves. So Shalabi recalls the pair reframing their thinking along these lines: “‘We’re at Google. Can we do things here that we’d never have been able to do as a startup?’ That’s when we were able to step back and say, ‘Let’s dream as big as we can.’”

As Shore explains it, their big realization was that most website owners—the ones who don’t have the expertise or resources to install fancy social networking features—have little idea who is on their site at a given time, what their users really like, what other sites they visit, and so on. If he and Shalabi could make the keys to that knowledge freely available, so that site owners could easily “turn the lights on the community that’s already visiting their sites,” that would be something the owners would really welcome.

With Shalabi doing the programming and Shore the product management and design, they decided to give it a shot. “Google is a culture which likes to really reward delivery,” says Shalabi. He calls it “rapid-fire, iterative development.” You put concepts out there, see what works and what doesn’t, you learn and try again—but if things look good, the company will step in with more resources.

“Our style is to fail fast, just get the stuff out there,” adds Shore.

In the case of GFC, they got a working prototype ready within weeks. By early 2008, Google assigned three more engineers to the effort. The team put out the private beta release that May and began working with site owners, bloggers, and others Google partners to try out new ideas. Shalabi says they had two goals. “One, we were trying to figure out if this was something people wanted. Two, if this was something that had value.”

Their conclusion, of course, would be yes on both counts. But early on it became apparent that one of their core premises—that people wanted to bring their friends with them to new sites—was wrong. “Initially, we had the thesis that you traveled the Web with your friends, and what we later discovered was it wasn’t really like that. It’s really about engaging with like-minded strangers,” Shalabi says. For instance, if you are a Formula 1 lover, it doesn’t mean all your friends are, he says. So rather than try to drag them to a Formula 1 enthusiasts’ site, what you really want to do is share your enthusiasm with others as passionate as you.

Until that point, they had focused on things like linking to users’ social networks so that their friends would know when they visited a new site. But, says Shalabi, “we discovered that wasn’t as relevant as saying something about yourself” to the community already on the site. So they retooled to pull in the user’s own personal profile and make it easy for GFC users to “declare themselves with a single click.”

Today, if you join a site with Google Friend Connect, your bio, photo, and any other profile details you’ve added are uploaded with a click from your Google, AIM, Yahoo, Netlog (it’s big in Europe), or OpenID account. You can later link your social networks such as Twitter, Orkut, and others to your login identity—and thereby pull in your friends to your new site, publish activities, and even do direct messaging back to the social networks. (“Facebook is not open in this regard yet. But we believe that at some point they will be,” says Shore). The “members gadget” installed on each GFC-friendly site is the core of Google Friend Connect. Other gadgets make it easy for the users to leave ratings, reviews, and comments and generally take part in the community on each site. “There’s a whole series of gadgets that help with that,” says Shalabi.

This is much like you’ll find on a social networking site such as Facebook, but the trick is you don’t have to do it anew for each site you join—and it is available to millions of websites, not just one. “It’s allowing you to add social functionality to any website,” says Shalabi. “Social isn’t just a destination, it’s everywhere.”

The public beta of Google Friend Connect debuted last December. And this April, Google really threw its weight behind the product by integrating it with Blogger, its free blog publishing system on which millions of blogs have been built. Every new blog created on Blogger now comes with the Google Friend Connect member’s gadget turned on by default. That, says Shalabi, has resulted in a “huge amount of traffic” for GFC.

I won’t try to capture all the developments since then. Suffice it to say that the team has increased the number of gadgets available. They have also built plug-ins to make it easier to add GFC’s social features to platforms such as WordPress and Drupal. GFC is now localized into something like 56 languages. And Google has opened the API (Application Programming Interface) to outside developers so that they can add features to GFC. The team, meanwhile, has expanded to 11 strong—10 engineers with Shalabi as tech lead manager, plus Shore as the product manager.

I also won’t try to run you through exactly how it all works. But I will say that after leaving the Plex, I tried out GFC both as a site visitor and a site owner, and it performed as advertised: two clicks to upload my personal profile and a paltry few questions before it was time to click the “generate code” button, which rendered the aforementioned code and this message: “copy and paste this code into your website where you’d like this gadget to appear.” Here’s a video about it:

“Eight million communities have formed with Friend Connect,” says Shore. Lots of them only have a small number of members. However, he says, smaller communities often “have the most potent exchanges.” And no matter what the membership size, if users are exchange valuable comments and information, “to me that means that content is being added to the web that otherwise wouldn’t have been there, and that’s more valuable to everyone.”

But how is it more valuable to Google? (I told you I’d get to this). I asked Shalabi about the business model, and got back a nice smile. “The economics of this are a little interesting,” is how he put it, which I took to mean either that he wasn’t exactly sure or that he wasn’t ready to talk about it. But a key thing about working at Google, he says, is that the goal is keeping Web users happy. “Our goal as a product team is to make our users the happiest users in the world. And if they’re happy, they’re on the Internet longer—and the economics just work for Google.”

Posted in Advices.

Tagged with .