A new take on an old adage

Enlarge

If it bleeds, it leads. The old newspaper adage tells us that gory headlines are good for sales. It is a tactic that tabloid newspapers have long exploited in their battle against dwindling circulation. Today’s news market is no longer determined by print sales, but by the range offered by online media and social networks. However, this has made the fight for attention all the more intense. There are statistics which show that the average time a user spends even on the main news portals is measured in minutes – per month. What does this mean for how media seek to draw in readers?

The old tabloid rule might feel like it belongs to the world of yesteryear, but is there not an equivalent rule in force today, even if less explicit? Researchers at LMU and from New York, Zurich, Stockholm, and Giessen have reframed the question in modern terms: Does it increase the number of clicks if headlines are studded with negative words – larded with outrage, as it were? Their answer is a resounding yes.

What is more, the researchers come to this conclusion based on their work with datasets from none other than the Upworthy.com portal, which started out with the mission to bring predominantly good news stories to its readers. “A nice little irony,” observes Professor Stefan Feuerriegel, Director of the Institute of AI in Management at LMU and one of the two lead authors of the study.

Sadness was good for clicks

For their analysis using methods from data science such as text mining, they were able to access a unique treasure trove of data from the Upworthy Research Archive. For years, the platform had experimented with its headlines and tried out different versions for their stories – a total of more than 100,000 headlines, which collectively generated around half a billion page views and almost six million clicks. After all, Upworthy.com was at that time one of the most frequently visited media portals, although the style that was so novel initially may have lost some of its shine since. But for half a decade, it registered more clicks than the website of the New York Times, explains Feuerriegel, and was the number-one fastest growing media site.

Click-through rates (clicks as a proportion of impressions) varied between zero and almost 15 percent. On average, the researchers ascertained, just a single negative buzzword could significantly increase the click-through rate, from around 1.4 to 2.3 percent. It should be noted here that these were rather mild words like “wrong,” “bad,” or “awful.” “Positive” words like “love,” “pretty,” and “beautiful” were less likely to engage readers. Longer cover lines – with potentially multiple negative words – increased the click-through rates. Greater complexity tended to be detrimental. When cover lines appealed to emotions, there were also some patterns to be observed: sadness was good for clicks, joy less so, while anger supplied astonishingly unambiguous results.

Depending on the topic, however, the effects of negative words had a different impact. The strongest effects were in the spheres of politics and the economy, but they were also very conspicuous in the categories of “people,” “school and education,” and “LGBT” – typical subjects covered by Upworthy – while the effects were less prominent in the realms of “entertainment” and “women’s rights and feminism.”

One of the things that makes the study so interesting, according to Feuerriegel, is that it does not look at the attention exhibited in comments and links, rather people’s private news consumption, highlighting which stories readers click on out of pure interest.

Collaboration instead of competition

Incidentally, the publication of the research is a story all by itself. Only when Feuerriegel and his team submitted an initial version of the paper to Nature Human Behaviour did they become aware that a research group led by Jay Van Bavel from New York University had also submitted a paper – in the same week, on the same topic, based on the same dataset from Upworthy.com, and with almost identical results. Instead of entering into competition, the two teams decided to collaborate on a joint paper. “That rounded off our results overall – not least due to the interdisciplinary cooperation between psychologists and computer scientists,” says Feuerriegel.

Claire E. Robertson, Nicolas Pröllochs, Kaoru Schwarzenegger, Philip Pärnamets, Jay. J. Van Bavel & Stefan Feuerriegel: Negativity drives online news consumption. Nature Human Behaviour, 2013