In 2010, Facebook introduced their famous Like button. Since that time, it’s found its way across the web. Back in 2011, celebrating the one year anniversary of the Like button, HubSpot published statistics on its use to that point outlining “2.5 million websites have integrated with Facebook (including more than 80 of comScore’s U.S. Top 100 websites and over half of comScore’s Global Top 100 websites)”. Fast forward to 2012 and the time of it’s S-1 filing, Yahoo! reported that Facebook was generating 2.7 billion likes and comments a day…31,250 every second! And that is only half the story, less than half. Even if 10% of Like button impressions were clicked (which is a stratospheric number), that means the “Like” button is still collecting data on the other 90% of traffic. Very rough math: 27 billions data collecting events from not only on Facebook but across the most popular sites on the web. ClickZ has a good article of the gold mine that is Facebook’s Like button and cite:
the company creates an impression log every time a signed in member views a Like button or social plug-in on a third party site, regardless of whether or not they interact with it. That record, which is tied to members’ unique user IDs, includes information on IP address, URL, date, time, and browser, and is retained for a period of 90 days.
This is an amazing accomplishment. Many in the industry including investors already know the power that this represents. Facebook has an engine that is learning with every page view about users and for certain, is synthesizing it into a data warehouse that marketers can only dream about. The possible use cases for such an engine are profound ranging from more astute targeting on Facebook with advertising and offers, to building an internet wide advertising platform (aka Beacon) that is every bit as powerful (if not more so) to Google Ad Sense.
Think about it from the perspective of a publisher on an online web site, for example, the New York Times or WSJ. They strive to learn about their audience and their behaviors so that they have generate more online revenue. However, they have an issue: their users have to register and be logged in so they can tie the profile of their user to their activity. Many publishers sit on a ton of anonymous cookie data that they cannot identify at a personalized level. This is why online behavioral targeting and ad exchange solutions are vital to them. In the case of the Like button, users have already registered at Facebook (you have to register to use the service) and thus an ID is in their system tied to an actual person. Each time the Like button is sent somewhere on the internet, Facebook knows who is looking at that page. A click on the Like button is just additional data. So in this scenario, Facebook most likely knows more about the users behavior than the publisher themselves.