Real-time social analytics: how editors become cyborgs

Successful online publishing relies on effective use of web analytics. In order to assess how a story will perform online and where it should be placed, most editors take input from Google Analytics or monitor trends in social media sharing. In many newsrooms there is a dedicated person who monitors statistics and makes recommendations to the editorial team on a frequent basis.

by WAN-IFRA Staff | March 11, 2013

The development of real-time analytics continues at a rapid pace, not only to speed up and improve the decision making process, but also to automate it. A new breed of suppliers combines newsroom understanding with expertise in data analysis. Strong relationships with Twitter, Facebook, Linkedin, et al are essential as sharing algorithms change on a daily basis.

Most U.S. publishers are already using real-time analytics tools from the new breed of suppliers, which includes Chartbeat and Visual Revenue (the latter very recently acquired by Outbrain). And CMS vendors are racing to integrate the tools into their own products.

To find out how paid content and editorial judgement intersect with real-time analytics, we caught up with Charlie Holbech, VP Operations and Co-founder, Visual Revenue, and Tony Haile, CEO, Chartbeat.

Our first question: Of all the tools and functions you make available to publishers, which one do they consider the most valuable?

For Haile, the most valuable function is helping “newsrooms to understand the pulse of their audience in real time”. If a story is underperforming, it can be acted on immediately. Haile focuses on Chartbeat’s proprietary metric, called “Engaged Time,” which tells an editor how well he/she is capturing the attention of the audience. He says, “If your goal is not just to chase traffic indiscriminately, but to build a loyal returning audience, it should be the key metric behind your decisions.”

Holbech talks similarly about benchmarking “which pieces of content are meeting expectations and which are not. Then we give them precise and polite recommendations on how to address the situation – immediately and at any moment of the day.” He emphasises the role of the editor, making decisions “not just by the data and our real-time predictive analysis, but by editorial voice and human judgment from the publisher itself.”

As more and more publishers move into paid content, how can social analytics support them?

Holbech is wary of relying on “vanity metrics” – simply checking re-tweets, favourites or replies. “This is fine for the individual, but not for media companies, where activity must be measured by whether it delivers and monetizes audience. […] But the editors in large newsrooms and leading editorial operations need to know more: How many views should this social content generate? How many views does it actually receive? Is this content that should be shared on social media? If so, when should it be shared to attain the maximum impact?”

Haile suggests two ways of using social feedback to improve a newsroom’s approach to paid content. Firstly, “Taking advantage of social spikes on ‘hunter’ content to move people towards valuable paid ‘farmer’ content is a key skill that requires understanding your data in real time.” Secondly, “Benchmarking the amount of Engaged Time for your paid content versus your standard content is a great way to understand not just how well you are driving traffic,” but also how well the division of content into free vs. paid is working.

Using analytics, some news sites position as many as 80 percent of stories automatically. How does this change the role of the editor?

Haile: “The editors’ role is to understand where automation can support them in areas where they’ve traditionally been burdened – the things they’re stuck doing, the time-sucks, that keep them away from making the best decisions for their audience.

“An editor’s job is to be focused on creating the highest quality content that will cause a reader to stay engaged and return. Anything else they’re forced to do – pull in related links or tweet a story that is receiving a high amount of social traffic – isn’t what they do best. It’s a waste of their time.”

Holbech: “Algorithms that determine automation can be improved upon when you inject distinct editorial viewpoints – not just most-read, most recent, or most shared stories. Our model, and we have seen it to be very successful […], is to apply a media property’s editorial guidelines to the algorithm. It results in a more optimised automation module for the publisher, and a more engaging, less commoditised experience for the audience.”

Above all, says Haile, it is important to avoid a short-term strategy that “compromises quality for what links are capturing the most clicks.”

Chartbeat customers include The New York Times, SwissCom, The Guardian, Gawker Media, and Forbes. Visual Revenue customers include The Sun, Le Monde, The Independent, London Evening Standard, and Sanoma Media.

Perhaps the two are not incompatible – The Boston Globe uses both!

See Tony Haile in person at Digital Media Europe, 15-17 April 2013.


Share via
Copy link