The Google Black Hole

Sergey and Larry just bought my company. Uh oh.

Late last month, Michael Arrington of TechCrunch reported that Google was in the final stages of negotiations to purchase Digg, the popular user-vote-counting news site. A few days later, Arrington, citing unnamed sources, posted an update: For unknown reasons, Google had walked away from the deal. Digg’s dalliance with Google wasn’t much of a surprise—just about every major tech company has flirted with buying Digg in the last few years—but to some of the company’s fans, the deal’s breakdown represented a real blow. “Google is really the only company I can think of that could actually improve Digg,” went one much-dugg comment.

The comment is revealing—among techies, selling to Google has never been considered selling out. In Silicon Valley, Google is seen as an entrepreneurs’ paradise, and not just because of the free food and fancy toilets. With its flat organization chart and the freedom employees get to work on creative side projects, Google is said to run much like a start-up: a company that heaps money and engineering resources on restless innovators who want to create the next big thing.

On the other hand, may I present Jaiku. Google purchased the “micro-blogging” company last year—think of it as a rival to Twitter—and then promptly closed it down to new users. Or look at JotSpot, a start-up that built a wiki collaboration tool for office workers. Google bought it in October 2006. Sixteen months later, it relaunched a radically different version of the service, Google Sites, to much criticism from longtime JotSpot users, who felt abandoned and betrayed.

Jaiku and JotSpot are examples of a phenomenon I call the Google black hole. Despite Google’s reputation for fostering new companies, many services that nestle into Mountain View’s welcoming bosom are never heard from again. The pattern: Company gets bought out. Users rejoice. Company lies fallow for months. Users grow impatient. Company’s employees get farmed out to other Google projects. Company lies fallow for more months. Users get even more impatient …

Consider the fate of Dodgeball, an innovative mobile service that was a predecessor to Twitter, Jaiku, and the many “location-aware” apps that now clog up the iPhone. The company, which launched in 2003, enabled people to send texts indicating where they were hanging out. In response, you’d get texts telling you which of your friends (or friends of friends) were nearby. In 2005, Dodgeball’s creators, Dennis Crowley and Alex Rainert, had just finished up grad school at NYU and were looking for investors in their service, which had become popular among techies in Manhattan. Of all the prospective offers they heard, Google’s seemed to come with the fewest strings. Google paid Dodgeball a small outlay of cash and stock—the exact terms weren’t disclosed—and Crowley and Rainert moved into Google’s New York office.

Immediately, Dodgeball’s founders saw that their corporate overlords didn’t want much to do with the acquisition. It took six months for Google to assign a single software engineer to Dodgeball. After a while, execs began pushing Crowley and Rainert to work on other things. The founders started to believe that Google had bought Dodgeball simply to acquire their savvy in the mobile social-networking business, not for the service they’d built. The pair quit Google in April 2007. The Dodgeball site is still alive, but no one runs it.

In many ways, Google’s Dodgeball acquisition was atypical. It was among the first after Google went public, and Crowley stressed to me that the amount of money that changed hands was very, very small. (Google declined to comment on Dodgeball.) But some aspects of the story seem applicable to other Google purchases. One difficulty Dodgeball faced was technical: It took time to move Dodgeball’s relatively simple codebase onto Google’s complex internal infrastructure—and after the transition, much of the system’s code became too complicated for Dodgeball’s founders to understand. Moreover, there were bureaucratic hurdles. As a start-up, Dodgeball had grown used to adding new features every week; under Google, the founders were told to get approval from layers of managers (though the Dodgeball crew began sneaking in small feature changes without alerting higher-ups).

Technical problems appear to be hampering other recent acquisitions. Last October, when Jaiku announced that it had been purchased by Google, the company’s founders said that they would close the site to new users “for the time being” in order to work with Google’s engineers. Three months later, new sign-ups were still down, and Jaiku offered another update on its blog: “To be honest, a lot of our time in the early going was spent on getting to know Google,” wrote co-founder Jyri Engeström. In April, Jaiku said again that its troubles were almost over. And then in May, after users complained that Jaiku was slow, Engeström promised, once more, that the service would ride high. “We feel the short term pain, too,” he wrote. “Thanks for sticking with us!” Now, 10 months after the acquisition, Jaiku still remains closed to new users. In that time, both Twitter (which is hampered by its own legendary tech problems) and FriendFeed, another here’s-what-I’m-doing start-up, have been signing up new people. Both, incidentally, were founded by former Googlers.

A Google spokesman assured me that Jaiku’s delays had nothing to do with the merger; he suggested that they were the normal troubles faced by any company trying to build a popular service. In an interview, David Lawee, the Google vice president in charge of acquisitions, outlined the rigorous steps Google takes after it buys companies. All prospective deals are approved by Google’s executive management team, and every merger is assigned an executive “sponsor” who marshals the resources—engineering, PR, sales, etc.—necessary to get the new company running within the Google infrastructure.

While Lawee acknowledged that it takes work to move a new company onto Google’s systems, he said that Google is “pretty accurate” at predicting how difficult that technical transition will be. Usually, Lawee said, the move takes three to six months, and its benefits are significant: YouTube, one of Google’s largest acquisitions, now slurps up 13 hours of video every minute, a scale that it would have had a tough time achieving on its own.

Lawee also pushed back against the idea of a Google black hole—even if the public doesn’t immediately see the results, the companies that Google acquires change the firm in big ways, he argued. Early in 2006, Google bought Measure Map, a much-beloved tool for bloggers to check traffic on their sites; the service has been closed to new users ever since. But Google didn’t buy Measure Map for Measure Map, Lawee said. It bought the company so it could add its features to another Google traffic program, Google Analytics—”and we did that.” Google had similar shape-shifting intentions for JotSpot, a company Lawee characterized as being “midway through a pretty exciting development plan.” JotSpot’s users might not like changes to the service, he said, but behind the scenes, JotSpot’s employees are hard at work throughout Google.

Considering Google’s history, Digg fans should probably be celebrating rather than lamenting the company’s nonacquisition. Diggers who were excited about a potential Diggoogle argued that such an arrangement would be preferable to Digg under Microsoft or another tech giant. But there’s little reason to believe that would be the case. TechCrunch’s Arrington reported that Google wanted to integrate Digg into Google News—not an indication that the Big G would have let Digg be Digg.

It’s important to note that acquisition hiccups aren’t unique to Google: In general, merging a start-up with a big company creates more problems than it solves. I called up Jason Fried, the head of 37 Signals, a successful Web company that makes software for small businesses and that has been adamant about staying independent. “You take great talents and you put them in this big company and they get drowned out by all this policy stuff,” Fried argues. “Putting a small company in a big company kills what was good about the small company.”

Is that what’s happening to Jaiku? Is it working on a supersecret social site that’ll blow away its competitors—or has it been left to wither? And what about GrandCentral, that great telephone company start-up Google bought a year ago—why is it still closed to new users? Nobody really knows. For now, all of those details are stuck inside the black hole.