Algorithms, metrics and analytics: the shift in how editorial decisions in newsrooms are being made

The School of Journalism's Nicole Blanchett was the lead author writing a chapter in  The Algorithmic Distribution of News (external link) , in which she shows how algorithms changed the way Canadian newsrooms run. 

Blanchett's chapter discusses algorithms used in newsrooms, including ones that help determine what kind of stories should be promoted and influence editorial decisions.

News organizations have long had access to audience data such as ratings and circulation rates. But what’s changed over the past few years is that news organizations now have immediate access to vast amounts of information through website data (external link) .

"So the real-time access to data in terms of who's looking at a story now, information about things like, for example, how far is someone reading the story? Are they actually consuming the whole story? Are they just dropping out after 15 seconds, all of that type of information," says Blanchett.

Algorithms and editorial decisions

With all that website data, Blanchett says editorial decisions can be made on how to engage readers, how to structure stories, and what type of headlines to use on websites.

This can go even further for news organizations that have their websites tied to revenue.

"If you are an organization and your revenue is tied to, for example, page impressions, which is something you sell to advertisers, in terms of this is how many times someone is going to basically look at this article and will see your ad, then that's going to change your process in terms of what is the most important aspect of how you're positioning this article," says Blanchett.

Blanchett says that there have been impacts on how audience data is perceived at an organizational level and how that impacts practice on the newsroom floor.

In the chapter, Blanchett talks about the Hamilton Spectator, one of the sites she had researched. 

"In 2017, there was a big push to basically find viral stories that could drive traffic on the homepage. Their internal objectives changed a bit. They were more interested in subscribers after this time that I spent in their newsroom. So what became more important in some ways was what type of articles can we have behind our paywall that will encourage people to become subscribers," says Blanchett.

Blanchett says this exemplifies how an organization’s goals influence how data are used to determine  what content is being promoted, created and privileged on news websites.

There can be a “shift from just trying to worry about views in the sense of 'how many people can we get to click on this' not really being worried about building a community, so to speak, versus what stories are going to be important enough to the community that's interested in reading our content, that they're actually going to subscribe and pay for the content," says Blanchett.

A shift in the industry

Blanchett’s interest in analytics began when she was on the journalism faculty at Sheridan College, meeting with the program advisory committee. Blanchett observed a shift in how people working in industry talked about and viewed metrics and analytics and how that was impacting process in many newsrooms.

"There was a much more open discussion about, yes, we want to use this audience data to make sure people are getting on stories and to maintain relevance and things like that. So that is what initially drew my interest," says Blanchett

Blanchett says the information from the book chapter came from two different projects: her Ph.D. research and data gathered for the Journalistic Role Performance project (external link) .

Audience engagement and community building

At Toronto Metropolitan University, specific courses, such as RTA912, explore the study of data concerning media audiences. Blanchett says that students need to better understand metrics and analytics within existing journalism courses.

"You don't necessarily need a course on metrics and analytics. What you need is for more courses to be talking about how metrics and analytics might impact the development of content within whatever type of content you're creating," says Blanchett.

Last fall Blanchett taught a journalism innovation course designed to teach students to understand who your audience is and how best to engage them.

"So at the beginning of the course, students are asked to decide, who's the audience that you're going to target? How do they like to receive information? And you can find that from a number of ways, like in terms of demographics of how many people are consuming content in certain ways," says Blanchett.

Blanchett also says that students explore how to find the best methods of audience engagement before even getting to the part of story creation and how to build a community with a particular audience.

"I think when you kind of combine that understanding of how do I best build a community with how do I best use the data available to me to help build that community? And then, how do I use those two principles? How do I create content that's going to engage and hit with this particular community? That's kind of the sweet spot of journalism," says Blanchett.

As for the future of algorithmic news, metrics and analytics, Blanchett says that they will continue to influence editorial decisions, especially since no specific business model works for all newsrooms.

"Different organizations are going to prioritize different things, depending on how they decide they can best make money and how they can best support the journalism that they want to do," says Blanchett.

The concern that Blanchett raises is the divide between what places have the money to invest in using technology, which organizations have cooperation between tech and editorial, and, those which have time to analyze data and make decisions.

"I have hope that these tools can be used well and more. More time will be spent in, again, bringing it all back to this idea of audience engagement; what works best for the audience? How can we best engage the audience on stories that are really important and that they need to know about? But my worry for the future is just that certain organizations will have the time and space to do it, and others will lag behind," says Blanchett.