I'm Feeling Lucky_ The Confessions of Google Employee Number 59 - Douglas Edwards [197]
In the United States, however, orkut lagged. Two weeks after its launch, a student at Harvard introduced a social network for his classmates. He called it "The Facebook." Within six years that service would have half a billion users. Orkut continued to struggle to secure a toehold. The issue, according to most engineers I spoke with, was orkut's inability to scale to handle the influx of traffic from an audience the size of Google's, a task it had never been designed to do.
Paul Bucheit, the creator of Gmail, disagreed. The real reason was "Google's tech snobbery getting in the way of its success." Paul said orkut "was taking off. Lots of people signed up. And then it got really slow." But that was a problem other social networks had experienced as well—MySpace and even Facebook ran into capacity issues almost from the beginning. The difference, according to Paul, was that those services jumped in and did whatever it took to make things work. Facebook was just a bunch of college kids. It had no brilliant coders like Jeff Dean or Sanjay Ghemawat. And the final configuration of MySpace, which had hundreds of millions of users, wasn't much more sophisticated than what orkut had at its start.
"But at Google," according to Paul, "that wasn't how you did things." Because orkut had been written using Microsoft tools, Google's engineers deemed it "not scalable." "They turned their noses up at it and they didn't make the thing work. They just let it die. And by the time they managed to rewrite it in a way that was acceptable to the engineers at Google, it was already dead everywhere except for Brazil. Who knows? If they had actually done what was necessary to make it go, it could have been successful."
To launch a radically new product from an established company, Paul asserted, you needed someone who not only believed in it but also was able to make the organization "do the right stuff." Two months after orkut's launch, he would personally put that philosophy to the acid test.
Bad News Arrives by Mail
I've never had much luck with email. For example, I had no idea that Microsoft Outlook had a two-gigabyte limit on how many messages I could save. I certainly didn't know that exceeding that amount would cause my inbox to explode and two years' worth of work to simply disappear. I found out, though, in 2002.
So I was receptive when Paul let me know he was working on an alternative to Outlook—a web-based email system he called Caribou—and asked if I'd like to try it out. I tried it, and it was pretty terrible. It didn't display well on my laptop, I couldn't sort messages from oldest to newest, and there was no way to select all the messages at once. Incredibly, there weren't even any folders for sorting mail by category. After a couple of weeks I told Paul, "Thanks, but no thanks," and went back to Microsoft Outlook.
A year later, I started hearing that Caribou had improved. Other Googlers were using it and not hating it, so I gave it another try. It still felt weak compared to Outlook, but it had some advantages. I could search through all my email quickly when I needed to find something, and it tied all my related messages together into an easily read thread. This time I stuck with it as Paul and a small team of engineers began prepping Caribou for launch as a Google product.
At the beginning of 2004, Yahoo, AOL, and Microsoft were the biggest players in online communication. They had created a balanced ecosystem of low expectations and commoditized email. Everyone knew web email came standard with a couple of megabytes of storage, inboxes littered with banner ads, and no easy way to find any message you had sent or received more than ten minutes earlier. Email addresses were disposable, and so many names had been claimed that almost everyone had to include a string of meaningless numbers in their user ID to open a new account. The major providers liked it that way and didn't want anyone rocking