Jul 14, 2011

My background working with social networks

Google's social network efforts hold a special place in my heart. I was engineer #2 on Google's first socal network orkut.com back in 2004, it was just Orkut, Adam Smith (the PM) and me for a while. Back then the engineers ran Google. They picked which projects got staffed by requesting to work on them. The orkut.com site launched because of one engineer, Orkut's, sheer will.. that and Adam Smith's PM skills and Marrissa's EMG support (Executive Management Group).

I got a phone call at home the weekend after the launch. Adam pleaded with me to help fix stability issues and get the site back on it's feet. I wasn't the best Google engineer for job. However, I was a Windows programmer and had just finished a successful run as the Toolbar 2.0 techlead, (I was responsible for the first intelligent popup blocker).

I knew how to scale client software and the required auto-updating and support across a wide range of client systems; however, Google infrastructure wasn't my strong point at the time. I was recruited because orkut.com was originally written in .Net and C#... that and the fact that nobody else on the Linux heavy engineering staff was volunteering to help.

I spent that first week getting the site back up … and the following year+ dealing with fires as we tried scale. It was rough and there wasn't very much glory in the job. Orkut (the person) had almost complete ownership of the product and it's direction. And while he listened to a broad range of feedback, ultimately, he had a vision and that's where the site headed. [Grammar note: 'orkut.com' is the project/site and Orkut is the person/engineer. We all have plenty of humorous stories around confusing the two references in conversations.]

Why was the site named orkut.com and not gSocial or something? Well, one goal was for Google to experiment with social networks. The theory went that we'd quickly iterate on different ideas and learn. Learn about .Net's scalability, learn about how users interact socially online, figure out what UI and features are core. Using the name orkut.com was an attempt to stay somewhat stealthy and not associate too closely with Google so as to control growth and expectations. In hindsight, this was very naive. All too quickly, orkut.com became huge... especially outside the US.

I've heard stories that in Brazil, when people figured out that Google and orkut.com were the same company, many thought that orkut.com bought Google. It's not unreasonable to say that orkut.com was the facebook of Brazil back in 2006 and Brazil was online in a big way. Much of this success is tied directly to Orkut's vision and dedication.

Because of this project, I did a lot of thinking about social networks. Gradually, the orkut.com team got more members and collectively we came up with experiments, ideas and theories about social interations. But the reality was that the team was understaffed and there were weekly fires to put out. At some point, there was such a large established user base that doing radical experiments wasn't viable.

So, in 2005, I started pitching a new social network project. One based on a secure, yet open model and would serve as a central ACL for all of Google's properties and anyone else who wanted to leverage it. In this early Google era, the way to get a project started was to recruit other engineers or get direct EMG support.

My pitch was: social networks aren't a product in-and-of-themselves, but they made all other products more compelling. Photo sharing, chatting, blogging, group discussions, etc all became significantly more valuable. A real social network would secure lists of friends like it protects other personal information. A user can have all their social clusters (aka circles) in a single site and insulate each group from the others. Having and keeping coworkers, family and burning man friends separate should NOT require multiple accounts.

I failed. All my documentation, presentations, conversations didn't result in a project. I blame myself for this failure, but maybe the time wasn't right.

Years later, I pulled out the documentation when Google started the latest social network effort and presented it to a few members of the new team. I considered joining the effort, but the team had a vision and personalities and I felt the last thing they needed was someone who had their own agenda (me).

I left Google back in 2009 after 7 years and I still miss many aspects of working there.


  1. Do you really mean that Brazilians thought orkut.com bought Google, or vice versa?

  2. The story I heard was that many Brazilians members thought orkut.com purchased Google.

    I should try to track down the source, but I can't even find any screen shots of the laser eye meme from 2004.

  3. I've heard the "orkut bought google" story too, or rather "google is run by orkut". I believe it.

    Nice write-up Tom!

  4. Thanks Patrick! Think you can dig up any old Laser Eyes screen shots for a different post I have?