Graph Search: Cultivating Big Data in Facebook's Walled Garden
Forget about any potential privacy issues with Graph Search. Facebook's new internal search feature faces bigger obstacles in that it is limited to only searching information provided by users, and won't work and play well with third-party search engines, namely Google. Those restrictions could keep Graph Search from being a Big Data game-changer.
01/29/13 5:00 AM PT
Facebook's recently announced beta of its new Graph Search resulted in the sorts of stories and headlines one has come to expect from a company whose every move is subject to media dissection. Some were skeptical about the value of the feature, especially its built-in restrictions: leveraging Facebook content alone and excluding well-established search entities like Google and Bing.
Others expressed concern about its privacy implications, since Graph Search seems likely to add further layers of complexity to what many consider an already overly-complicated and difficult to parse privacy system.
Surprisingly, there were few commentaries on how Graph Search might fit into the Big Data projects currently being considered or pursued by numerous organizations across the IT industry. Why was that a surprise? Because on the surface, Graph Search has obvious Big Data implications for Facebook, its community of users and its customers, the myriad businesses that purchase advertising and other services from the company.
Using Big Data to Ease Big Pains
By definition, the sheer number of Facebook users, and the volume and variety of information they share and exchange, would make any dedicated search solution a Big Data play. This is also implied by the way the company is positioning Graph Search as a feature that will allow users to search "for more than you've been able to find before."
That brings it into line with the practical benefits Big Data vendors offer clients overwhelmed by the rapidly expanding information assets their organizations generate and save.
An often unspoken Big Data goal is to create a framework or mechanism that helps bring unwieldy information assets into better order. Why? To improve current access, ease future management requirements and enhance the long-term value of data investments. In fact, I expect behind the scenes work concerning Graph Search should help Facebook capture increased efficiencies in its storage, data center and information management efforts, which will significantly impact its bottom line.
Online Community Value and Valuation
Since the beginning of the commercial Internet, businesses have attempted to build communities of consumers that they could then exclusively exploit. That was certainly the case with early ISPs and e-commerce wannabes. But most of those players eventually ran up against the Internet's great leveling ability, which made it as easy to lose customers as to gain them. If consumers dislike a site's features or find them too restrictive, they can easily find another service that meets their needs.
By this measure, Graph Search has a potentially dangerous "back to the future" vibe about it. Facebook's decision to limit searches to its own information, to exclude other search results and to prevent other search engines from collecting Graph Search data is clearly a "walled garden" of the concrete sort. What's to keep potentially disgruntled Facebook users from heading to the exits, like those who abandoned AOL and other stalwarts of the dot-com boom?
First and likeliest is scope: no other social network has attracted Facebook's one billion global users. Despite Graph Search's potential to be a distinctly isolated and isolating online experience, the company appears to believe the vast majority of users will stay because they have essentially nowhere else to go. That's a big bet, but since Facebook could lose tens of millions, or even hundreds of millions of users, and still remain the largest social networking community on the planet, maybe it's also a smart one.
Cui Bono? Who Profits?
Over time, most successful Web-based businesses have innovatively leveraged data broadly available through the public Internet, along with proprietary information collected from client interactions. Amazon is a good example of how this approach can be used to create a self-perpetuating and profitable business model.
Though Facebook hasn't discussed its commercialization plans for Graph Search, it seems safe to assume that the company could leverage the service to deepen understanding of its users, and then sell that knowledge to advertisers and other core customers. In addition, if Graph Search performs as the company hopes, plans to extend the feature to the Facebook mobile platform later in 2013 will likely impact and even underpin the company's mobile ad strategy and offerings.
There are two unknown quantities here: if Facebook users will actually adopt the new service, and how accurately their usage reflects their personal behavior and real world interactions. There are some similarities between what Facebook intends with Graph Search and the work of early pioneers of landscape design such as Frederick Law Olmsted. In urban parks, a critical point is the illusion of equivalency, or how well an architected construct can evoke the natural world.
Facebook appears to believe its billion-strong online community constitutes a similar equivalency with the wider world, enough so that its users will happily embrace Graph Search and abandon deeper search offerings like Google and Bing. What the company appears to have missed or ignored is that, however large and well-appointed, public parks are meant to reflect the natural world, not replace it.
Overall, I find Graph Search curiously uninspiring and pedestrian. Considering the project's scope, its extreme technical requirements and the size of Facebook's information assets, Graph Search could qualify as one of the most ambitious Big Data projects of the decade. But the relatively mundane benefits it promises users and the limitations of worldview required for adoption seem somewhat out of balance with the effort expended.
It's as if a country built a state-of-the-art power plant to provide sustenance to its populace, then fed them nothing but corn flakes and bologna sandwiches. More practically, Graph Search does draw a firmer, more obvious line between Facebook's users and its actual customers. In this case, users are being asked to adopt a more limited and potentially limiting search methodology in order to improve the advertising and other services that Facebook customers buy.
Like every walled garden before it, Facebook's utopian conceit is that its artificially constructed community is large and various enough to meet all the needs of its members. That is a wager that virtually no community or company has ever won. If the past is any predictor, Graph Search will win a few stalwart fans and numerous casual users. But we expect that many more will eventually recognize Facebook's walls for the impediments they are and, instead of enjoying the neatly trimmed grounds, simply walk away.