At Google, Missteps Are A Key Part Of InnovationAt Google, Missteps Are A Key Part Of Innovation

Sometimes, when you're trying to innovate, your best intentions can be misinterpreted. And when you're Google, and your phenomenal success results in a growing "kick me" sign on your back, you have to take this into account. So would be the lesson Google should take from a note information received from an alert reader recently in response to our coverage of Google Transit, the latest experiment to emerge from the combination of Google Labs and the company's famous policy encouraging employees to spend 20% of their time pursuing crazy ideas.Google Transit ambitiously seeks to make the world easier to navigate by giving us a tool for planning trips on local transit systems. So, instead of going to separate sites to find out local schedules for trains, buses, and light rail systems, you'd enter your starting point and destination, and presto, you've got a set of directions that tell you where to go and when in order to get from point A to point B as efficiently as possible. Currently, the tool, which is a beta test, only provides these capabilities for Portland, Ore., and not all that well (more on that in a moment).

But the focus on Portland is what led to the aforementioned misinterpretation. The reader wrote to us suggesting that Google co-opted the data from a similar tool built by the Tri-County Metropolitan Transportation District of Oregon (known as TriMet). And at first glance, it certainly looks like a distinct possibility. But as it turns out, Google's focus on Portland was a carefully thought-out and very smart move: TriMet actually agreed to share its data with Google, which identified the agency as being on the cutting edge of using technology to make its system more usable for riders.

Here's where Google erred, though: The reader would have known the background if only Google would provide an obvious link on the Google Transit site to an informative Google Blog entry (scroll down to the Dec. 7 post) that explains the project quite clearly. Note to Google: Don't let such helpful blog postings go unread. Promote them in the right context so users can benefit from their insights.

As for the performance of Google Transit itself, well, let's just say it has a long way to go to match the effectiveness TriMet achieved. Whereas TriMet's tool recognizes just about any Portland reference you enter, from addresses to landmarks to neighborhoods to transit stations, Google Transit really struggles. I couldn't get it to bring up a host of street addresses I know exist in Portland, and it inexusably doesn't register names of transit stations, seemingly an important detail if you're trying to navigate local transit. In one case, through sheer persistance, I did get it to recognize an address on Ellis Street--unfortunately, the Ellis Street it found was in the United Kingdom, which at last check wasn't part of Portland's transit system.

But this is nit-picking, because the fact that Google Transit appears to be somewhat buggy is part of what makes it beautiful: Unlike most tech companies that fret over every product decision because God forbid we should invest in anything that might not make us money, Google understands that true innovation requires a lot of failures along with the successes. As a Google spokewoman told me, "There's no guarantee that our experiments will turn into products." And while Google Transit can hardly be labled a failure so early in its history, at least Google is willing to stumble publicly in its pursuit of the next killer tool. That, dear readers, is a pioneering spirit that's all too rare among today's crop of tech vendors.