When it comes to using Big Data, Facebook occupies rarified air along with Amazon, Apple, Netflix, and Google. It's a point that I've made countless times before in my talks, books, and blog posts. But does that mean that the company has perfected its use of vast troves of mostly unstructured data?
Hardly.
The company's struggle with its iconic news feed feature has been well documented. Long story short: Facebook still doesn't know what types of news and status updates to show its users. One solution: Gather data–i.e., conduct experiments on its users, although Facebook didn't expect to face a considerable level of public outrage.
Ethical or not, Facebook has to eventually solve its burgeoning relevance problem–or at least minimize it. Unengaged users are a massive threat to the site’s business model. Remember that users aren't customers. Facebook must monetize its users. If user "engagement" begins to wane, advertisers will soon take their business elsewhere. The house of cards comes down.
Beyond Experimentation
Of course, experiments are just one club in the bag. Like Google, Facebook constantly tweaks its algorithm and tries to match stories based on what it believes to be relevant keywords. Case in point: Here's a series of related articles that Facebook automatically generated based on a status update I shared. (The post and link concerned Amazon's dispute with publisher Hachette over e-book pricing. I asked the question: Are books special?)
Note Facebook's second related story on a swimmer in Australia. Is it related to Jeff Bezos' crusade to lower e-book prices? Not even close, but Matthew Richelle's tragic death shares one commonality with that story: the name Hachette. Yes, Facebook equated a town in Australia with a publisher currently embroiled in a dispute with the everything story.
Simon Says: even Facebook struggles with big data
Although Facebook almost certainly deals with more data than your organization does, it also puts more data-related financial and human resources to work. Make no mistake, though: The company hasn't solved or perfected big data. Know that going in. Big data is a marathon, not a sprint.
Feedback
What say you?
