Nate Westheimer's observation about Google's App Engine: Aiming At Facebook, Not Amazon set me thinking. Should Facebook's F8 platform be considered a contender in the platform-as-a-service wars?
Since my perspective is firmly on PaaS in the business world rather than as a consumer play, Facebook isn't really on my radar (I've always firmly resisted any passing urges to set up an account), but that doesn't mean it should automatically be discounted. After all, many individuals and some organizations do use Facebook for business purposes. Most famously, Serena Software, which last year adopted Facebook as its intranet. I was somewhat skeptical when I first heard this but the company's SVP Rene Bonvanie assured me when we met a few weeks ago that 740 out of 820 employees are active users, which is a lot better participation than Serena's former intranet ever managed. The Facebook platform of course constrains applications into a social networking framework, but that's no different from the functional constraints imposed by a lot of other PaaS application builders (the topmost of my five-layer categorization of PaaS). Marc Andreessen's Ning is another platform in the same mold.
The problem for any business considering Facebook is that it's a determinedly consumer play, to the extent that I don't think it can ever seriously fly in the enterprise. Facebook trades free functionality in return for attention and relationship data — and users give up a lot of their control over that data. Businesses aren't willing to make that trade-off. Bonvanie shrugged when I put this point to him, saying that very little of what Serena's employees post to the Facebook intranet is proprietary to the company, but at the same time he admitted that Serena stores its company confidential documents elsewhere than Facebook, so its role as an intranet platform is more limited than you would typically look for in an enterprise setting.
Where Facebook will suffer in competition with Google App Engine is in developer mindshare. The App Engine is going to take a lot of the shine off Facebook as the hot new platform for budding developers to try next [update: Niall Kennedy has posted a useful developer's-eye assessment of App Engine]. It could be an especially potent competitor if Google incorporates support for OpenSocial into its identity infrastructure, as outlined by David Recordon.
Whether App Engine will permanently challenge Facebook, though, will depend on one burning question: will developers be able to make money from their Google App Engine creations? Make no mistake about it, developers build on Facebook not only for the kudos but also because a successful widget can make them a lot of money. Rod Boothby of Joyent, which offers free hosting for Facebook developers, says that some of its customers have built apps that earn thousands of dollars a month (of course once the app scales to a certain size, developers have to upgrade their account, so Joyent gets payback for its pump-priming). It's the hope of building a successful application — one that results in a new income stream, a better job or even venture funding or an acquisition — that drives so many to invest time and effort into experimenting with these new platforms. Every developer would love to build the next Microsoft, Netscape, Google or YouTube.
If Google App Engine can do a better job than Facebook of appealing to the business market, then it opens up a rich new seam of entrepreneurial opportunity. Businesses pay real cash for useful functionality, out of a much bigger pot than you'll ever reach by taking a cut of advertising spend. But for partners to be able to tap into that potential revenue stream, Google will have to finally work out how to bill for usage (and share that capability with their App Engine developers).
I'm old enough (and more) to recall the excitement when the Google API launched back in 2002 and people first started discovering what could be achieved from building Web mashups. Google stopped offering that SOAP-based API in December 2006 in favor of an AJAX API that shows ads alongside the search results. I described the decision at the time as a retreat back to Web 1.0:
"Put bluntly, it confirms Google's abject failure to monetize its API except by the indirect mechanism of contextual advertising ... Google's track record is in stark contrast to Amazon, which has had several APIs emerge from (sometimes long-running) beta programs as pay-per-use commercial propositions, and is now building a serious business around those APIs, serving an expanding community of developers."
The App Engine creates an opportunity for Google to recapture all the original excitement of the Google API back to its platform. What it must do this time, though, is go beyond merely competing with Facebook for developer mindshare in the consumer market for platform-as-a-service. It must also compete in the business market with all the many PaaS competitors emerging there — everyone from Salesforce.com to Zoho (read this great post by Zoho CEO Sridhar Vembu for further insight on segmenting the PaaS continuum). But to succeed there, it will have to give its developers a better business model than just selling ads.
So to sum up. Yes, Google App Engine is a big competitive threat to Facebook as a platform for developers who want to reach out to the volume consumer market, especially if App Engine developers can build OpenSocial functionality on top of App Engine. But Google also has the opportunity to compete in the business market, where the volumes per application are lower, but the potential revenues per customer are massively higher — provided it makes sure that App Engine developers have the tools to monetize their applications.
Footnote: By the way, I'd like to thank Google for the unconscious (and no doubt unwitting) reference to the title of an analyst report I wrote back at the end of 2000, Internet Application Engines. Sadly it's no longer available from the publisher, but this column from October 2000, When the Internet Becomes the Computer ..., gives a taste of its now eerily prescient themes (emphasis added):
"Instead of the discrete islands of conventional enterprise computing, Internet computing thrives in the form of multiple collaborative components ... Each individual piece of computing becomes a component service — an Internet application engine — within the interconnected global infrastructure. The task that's now under way is the creation of an Internet operating system that allows those many components to interact seamlessly within the infrastructure. When the process is complete, the Internet itself becomes a shared global platform for the automation of commerce — along with every other form of collaborative human endeavour — built from millions of participating components and services."