Bit.ly’s chief scientist Hilary Mason was the featured speaker at the PW and TOC Publishing Executive Roundtable on May 22, and the key topic was “big data.” Speaking to a room of publishers, Mason said publishing’s shift to a more dynamic information model is “inevitable,” and that, at this point, dynamic data is already being generated, it’s just about optimizing it.
As a chief scientist (a position that is growing in multiple areas, including the government) at Bit.ly, Mason looks at what people share and read and interprets that data, and how ideas proliferate online. When it comes to publishing-related content, there is typically “a burst of attention and a beautiful decay,” according to Mason. To illustrate this burst and slow decay (which tends to be slower for evergreen content), Mason said that a discrete piece of content’s half-life on Twitter is 2.8 hours—meaning half of that piece of content’s clicks will happen in its first 2.8 hours. Other social media half-lifes: 3.1 hours for Facebook, 7 hours for YouTube, 7 hours for StumbleUpon. Mason also outlined the peak traffic windows for these sites: Twitter aligns closely along business hours, Facebook reaches its peak in the evening/down time, Pinterest peaks on Saturday mornings, and Tumblr peaks late at night.
Speaking about the psychology of the average Web user, Mason said, “People share things that make them look good,” meaning that they share breaking news, or content that makes them thoughtful and perceptive. And for companies, Mason stated that the right audience is better than a big audience, and that engaging with the right people breeds more successful and more engaging content.
The pull of publishing (along with many other sectors) toward harnessing and understanding “big data” certainly seems like the direction in which the industry is inevitably going, and all in attendance agreed that developing the data in data-driven decisions is going to be of high importance going forward.