We have been in contact with Yury Lifshits, the author of the Yahoo! Labs Ediscope paper, who is now in Russia working on not one but four new non-Yahoo projects. Unfortunately none of those involve social analytics, so our earlier hypothesis that an excess of tech investment money caused Ediscope to shrivel on the vine may not have been correct.
Now, this contact was by email, and shares the pallid character of all such “interviews”: bloodless interactions lacking a pulse. The biggest problem with email interviews, on a practical level, is that you miss the follow up question and if the respondent ignores your point – Yury I’m looking at you – you can’t set him straight immediately. Follow-up emails are tedious.
So I’ll cut to the good stuff: on the question of “what makes a story shareworthy”, which Lifshits identified as the “main question”, he responded with four points.
- Direct relation to the reader’s life
- Appeal to your core audience – the “power readers”
- “Discussion-ability” – The ease with which people can continue a conversation having read the article
- “What the fuck?” feeling
The list is fascinating for the Share Wars project. What’s great about it is that Lifshits is thinking entirely about the way an audience responds to content, rather than anything inherent in the content itself. His instinct is to abstract to the level of audience response. “Look for content that does this to your audience”. It is an approach similar to Berger and Milkman’s in their seminal “Virality” study.
We can dice it a number of ways. The first point, a relation to the reader’s life, is shared by most classic theories of what makes a good news story. In those theories, as in its relationship to “shareworthiness”, it is a sufficient but not necessary factor: there are plenty of good and shareable news stories that have no direct relation to the reader.
The second point is more prescriptive than explanatory. The massively overrepresented contribution of a core of power users to site traffic has recently been called out by the Pew Research Center (Navigating News Online – one of our foundation links). The assumption is that you could focus more on this core group to increase traffic with great efficiency: increase your power users an inch and move traffic a mile. But the assumption has not been tested – who knows if in the ecosystem of users the power users are not a necessarily small proportion of the whole? Like sharks in the maritime food chain, not every fish can be a super-predator. You need a fair whack of sardines.
There is a peripheral point behind the idea of a particular group contributing more – that shareworthiness is not an entirely universal set of attributes but will change according to audience. Different types of stories will be shared more by different groups. One of the aspects of sharing that Share Wars is dedicated to researching is the difference in shareworthiness from country to country and between publications.
Lifshits’ third point is interesting. A story that encourages or facilitates a conversation. A story that can be used by the reader. A story that can be retold, agreed with, disagreed with, picked over. A memorable story. A story that encourages its own transmission.
We are not going to begin a discussion of memes, or self-replicating ideas, at the moment, but there is no doubt that a tendency to encourage transmission is relevant to a story’s share-ability. For example, stories that concern Facebook are over-represent in our lists of most shared stories. These stories are talking about the very medium in which they are swimming, which creates a feedback loop. James Gleick makes the point that all information transmission technologies have the natural advantage of facilitating their own propagation – one of the ideas they spread between people is the idea of themselves.
Based on this principle, stories that flatter the receiver/transmitter should also be overrepresented. We will see when we look through the data.
The final point in the Lifshits list: what the fuck? We agree. As we heard from Gleick, information implies surprise. The nearer the surprise quotient to the WTF constant, the greater the shareability. In a gentler age, it would have been “amazement”.
Lifshits has swapped Silicon Valley for Saint Petersberg and social analytics for educational projects. But he still believes in the power of sharing:
“On average, news sites are getting 1% of traffic from social networks. But smarter sites are getting double digits. The share of social traffic is growing. There is more sharing than is directly measured by analytics systems. Sharing is a feedback mechanism. You start to write more of the stuff that is shared well.”