The Science of Facebook Likes, the Viral Web, and Charlie Sheen: a Q+A With Yury Lifshits

A few weeks ago, Yury Lifshits, a young scientist at the Web-obsessed data den Yahoo! Labs, quietly published the Like Log Study, one of the most in-depth statistical examinations of the ways our media diet has been shaped by that peculiar and ubiquitous little "Like" blue button you see above and on nearly every web page. Lifshits, who was born in St. Petersburg and studied mathematics at CalTech, sheds light on what looks like a fundamentally shallow online readership, as evidenced by the top five headlines from the Anglophone internet during the period he studied:

• Why Chinese Mothers Are Superior: (342,294 – Wall Street Journal)
• New Zodiac Sign Dates Earth Rotation Changes Horoscope Signs (328,192 – Time Magazine)
• No, Your Zodiac Sign Hasn't Changed (238,596 – CNN)
• A Holiday Message from Ricky Gervais: Why I'm An Atheist (164,166 – Wall Street Journal)
• Tired Gay Succumbs to Dix in 200 Meters (152,780 – Reuters)

Some of the interesting lessons from Yury's numerical analysis, as I wrote when the study was released, shouldn't come as surprises: we tend to use the web to share things that we can relate to more often than we want to read about far flung people and news, and we tend to "Like" the sort of content that will make us, well, likable.

I caught up with Lifshits over the phone to get a better understanding of what he found, what he likes, and – hold your horses, all you social media wizard-wannabes – what the secret is to total viral internet domination.

Motherboard: Hi Yury. Why are "likes" important?

Yury Lifshits: Journalism needs feedback. In print journalism, there was no granular feedback on individual articles. It was in the form of letters that would reach editors months later. It was private, even on the web until recently. The New York Times knew how they were doing, but not how another publication like the Wall Street Journal was doing. With the likes, you get a public signal—external people can judge them now. Everything becomes transparent. You can see where the demand is and where the supply is and figure out where they should meet. Because it is measurable, it means the speed of learning will grow dramatically. People will improve and change their editorial policy not once every week or once every decade, but once every day.

The second big discussion is whether you should respect quantum demand. Essentially, there were two schools of thought. One school says the editor is the king. The editor knows what to write much more than the reader knows what he wants to read. The editor is smarter, like a Steve Jobs. This is an old school of thought.

The new school says, "forget that, let's see what people are searching, and lets create content for that. Whatever advertisers are ready to pay for we will just write about that". Now with likes you can capture this implicit demand—what people really share a lot, you can uncover this demand, and you can create content directly for this demand.

What kind of content are people demanding?
There is a very clear signal of what people are demanding. People don't care about anything related to how content was produced—it's just not important. People don't care if you paid, if it came to you for free, or if it was sponsored. Just the fact that you broke news 10 mins before CNN doesn't give you much advantage. The old school thought is that faster and more exclusive reporting leads to success.

The social media winner is the media source that publishes something directly relevant to a normal person's life. Not the White House, not Afghanistan. It's got to be news about the reader. I want to read about myself.

You have to think and imagine how someone will tell other people about what they just read. To say "something happened" is rarely interesting enough to tell be worth sharing. Writers need to disagree with something, or to write about something that will be useful in their readers lives tomorrow.

Charlie Sheen. What the hell?
I have not followed this in depth, but I totally understand why the Middle East situation is not engaging. In my studies, stories about Iran and such are not engaging. If you live in Chicago and today you feel OK and tomorrow you feel OK the next day and something minor happens in Afghanistan, you're not about to call your mother about it. Just because something is important in human history, doesn't automatically make it something people are going to share.

Somehow Charlie Sheen is also about people themselves, but I have not deconstructed it yet. I think Charlie Sheen is about the dynamics of life—how you fall down, get back up, go crazy—maybe everyone knows someone crazy around them.

Then the other issue is that the media follows "rich get richer" dynamics. Media always needs a subject #1, and they will always write a bunch of stories about this subject. Sheen became #1 for this month, but Wikileaks was this story a few months ago. It faded away, and now we have Charlie Sheen.

The metrics bias journalism towards quantity. How focused do you think journalists should be on accruing Facebook Likes?
This is easy to answer, but the answer is complex. First, there is a little but significant part of journalism which is about public service. Finding scoops, bringing attention to social problems and things that need to be fixed, applying pressure to get results—this is journalism as public service. If journalists think of themselves as public servants then they shouldn't worry about Facebook Likes—they should focus on transparency and stopping things from getting worse.

The other side is entertainment, education, and information. In this part, there is greater flexibility in what gets written about. How to be successful in dating; how to eat well; how to exercise; how to play guitar; how to iPad DJ. Here, it's only good if you get the right signals and figure out what to write about. When you speak about signals and recommendations from the reader there are not just one but two recommendations: topic and form.

I think that writers now do a reasonably good job of picking topics, but they are writing in the wrong format. Sometimes they write text where people want lists, facts where people expect images or video. How much should you post? Once a day? A week? 100 tweets and one article, or five tweets and two articles? How you balance formats is how you can manage readership.

It seems like most of the sharing is happening amongst a small group of people who have an outsized influence.
I did not mention this in the study, but I am now realizing how important it is to realize that stories are not shared by random people. It is a special, predetermined community of people who have a large following themselves and who enjoy passing things around. They have their own blog or whatever and love to pass things to their followers. The point is not to please the regular reader, it is to create a story that is unique enough that this retranslation network will take notice of it and pass it around.

It comes down to catching the attention of people who have large followings. There is only a small threshold between the average story and the thing that gets picked up by the New York Times. A big story is not a story that regular readers like a little more, it is something that is picked up by people who can start epidemics – it's a matter of critical mass. Every month, writers need to put in the work to make a possible virus a real virus.

Check out Yury's Study, follow Yury on Twitter, and Like our coverage of his study on Facebook.

PHOTO: holt_lyda/ Flickr