Site Meter

March 14, 2010

Gmail Failures, Crazy Ideas and Wave's Leapfrog

On the Web, there is a lot of confusion over where Google Wave fits alongside the recently-introduced Google Buzz, or even if Wave is supposed to be a companion or competitor to Gmail - which could potentially cannibalize the company's extremely visible (and profitable) e-mail business. Today at the SXSW conference, the team leading Gmail said the company has to take risks, learn from mistakes, and yes, even sometimes build products that are in conflict and may replace one another - in the name of keeping competition from doing it themselves.

As it was described this afternoon, Google Wave, which debuted in early beta last year, is a "leapfrog project", which goes beyond today's environment, but is set to impact a future Web. The team working on Wave, as discussed with the product launched, is looking to do more than just build a collaborative service, but to possibly even replace e-mail itself, something the GMail team recognized might seem at conflict to their core mission.

"When people ask if we are cannibalizing our own services, we would rather cannibalize our own services than have other people do it," said Todd Jackson, product manager for Gmail and Google Buzz.

Putting significant resources into disparate product lines that may come to future conflict might seem crazy, or even a bit paranoid, but it sounds like that is par for the course for the team, which said it likes to take big risks, which might not ever see the time of day, or die when they do. In fact, we learned today that Google Buzz's original incarnation began several years ago.

"Most of the things we try fail," said Jonathan Perlow, software engineer on the Gmail front end, responsible for GMail Chat and Mail Goggles. "We have lots of things that are false starts. We recently launched Google Buzz, and it had some false starts before it launched. We started something like Buzz around when we launched chat four years ago. Good ideas live on, and you figure it out."

Figuring things out for the Gmail team can be very quick. The team boasted of a close-knit engineering environment where ideas can be discussed and coded quickly, and where meetings are the exception rather than regular practice.

But while most failures for Gmail have occurred in testing and not made it outside of the walls of Mountain View, the initial failures for Buzz happened thanks to the team making some core mistakes and not having a testbed of real-world users, relying too heavily on their open corporate mentality.

"Gmail thought that e-mail and chat networks were also the social networks, and we missed the boat there," said Jackson. "(Autofollow) worked really well within Google in a trusted environment. Googlers rarely used block."

While Gmail's focus has changed over the last six years with the additions of Chat, Buzz, user interface updates and other features, the product initially aimed with three main goals: enable users to never delete e-mail, have a spam filter that really works, and build a Web interface with the level of quality of a desktop application - concepts that nobody knew how to do, but wanted to accomplish anyway.

"One of the lessons I learned is that when we start with ideas that are crazy at the time, but we thought we could do, they would be pretty great for users, said Perlow. "They had no idea how to build these things, but had to figure it out."