Who has more clout in spreading the news: the New York Times, the Guardian or Wired? Such questions have been the stuff of cocktail chatter but now, thanks to the rise of Twitter and big data analytics, we have some hard evidence.
- What The Guardian really thinks about newsprint
- Content newswire NewsCred buys cloud publisher Daylife
- Pearson sorry for downing 1.45 million teaching blogs
In a new study, two University of Arizona researchers use Twitter’s emergence as a “serious newswire” to compare the reach and longevity of news stories tweeted by organizations like Reuters, NPR and the Washington Post. Over a three-week period last winter, the researchers looked at tweets containing story links and found that stories from the BBC and the New York Times were the most widely retweeted.
The study’s authors, Sudha Ram and Devi Bhattachary, also looked at metrics like articles’ half-life to determine the popularity and longevity of a news story. They found that articles from BBC, Mashable and the NYT had the longest life span, while the BBC, Mashable and Wired were most likely to publish popular articles — stories on Twitter that exceeded the average article half-life of 5.5 hours (“half-life” is based on a bitly definition that says it’s the amount of time at which a link receives half of the clicks it will ever receive after it’s reached its peak).
The study also looked at rates of engagement — how often a Twitter user is likely to tweet a given news source. On this front, financial publications like the FT and Forbes scored lowest while the NYT, NPR and the BBC scored highest.
The BBC’s prominence can be explained in large part by the fact that is has three major Twitter spigots that frequently retweet each other: “bbcnews,” “bbcbreaking” and “bbcworld.” This means that the BBC has far more of what the study calls “Maximum Level” retweets — Level I is an initial retweet, Level II is a retweet of Level I and so on.
The study, which draws on methods used for epidemics and network analysis, also uses intriguing graphics to display news organizations’ influence. This picture, for example, shows how the NYT and the Washington Post stories produce similar network effects, but the NYT stories are retweeted by more people in isolation:
So what to make of all this? One obvious observation is that the pool of data tied to Twitter gives news agencies unprecedented tools to measure their influence and shape strategy. But, as the study notes, one size may not fit all:
This leads to the question of what constitutes successful news diffusion on Twitter. Bursts of 1st level tweets within the first hour of diffusion (corresponding to instant reach to a large audience) or a high network diameter indicating multiple levels of exchange of news over a period of time (longer lifespan)? This depends on the objective of the news media source.
Another takeaway is that these are still early days for data and news analysis. While the Twitter study is intriguing, it is presented (appropriately) in the language of science — “edge/node ratios,” “ego network details” and so on. This means it may take time for the study’s implications to be translated into everyday guidance for publishers and editors.