The Pageview is Passé: New Metrics Emerge to Measure Audiences
revenue-wordcloud
Credit:
March 18, 2014


Editor’s Note: This is the first post in a series examining the influence of analytics on how news is produced. 

Upworthy, a web news aggregator known for attention-grabbing headlines and viral videos, recently announced a new traffic metric it calls “attention minutes,” reflecting how much time readers spend on the site overall and in each article.

Chartbeat, a web traffic monitoring service, offers a similar metric called “engaged time,” which measures how much time people spend “actively interacting” with articles by doing things like scrolling up and down the page.

ESPN too has its own special metric, called “average minute audience,” which calculates the number of users per minute, according to Digiday, which notes the benefits of this marker; it can capture the number of users during a particular news event and can easily distinguish popularity between devices.

The pageview is fast becoming passé, a number that is easily artificially boosted by photo galleries, slideshows and other tricks. Instead of turning to this metric, news outlets  are seeking better ways to measure reader behavior, combining pageviews with other measures of traffic to get a more comprehensive view of what’s happening on their sites. Some are building custom dashboards, such as the “Ophan” system developed by a hackathon team at the Guardian.


Follow AJR on Twitter: @AmJourReview


All of these tools reflect the struggle editors and business managers face as they try to figure out which metrics matter most– and whether they should be allowed to influence editorial decisions.

“We give a lot of thought to what exactly it is that we’re trying to do,” said Daniel Mintz, Upworthy’s  head of analytics, via email. “And we believe deeply in the fact that you are what you measure. So, since our goal is to draw massive amounts of attention to things that matter, we thought it was worth trying to measure that directly.”

The ability to see and count what readers do on a website provides a relatively  new source of audience intelligence to the news industry, which historically has relied on self-reporting through surveys to gauge what people actually read.  The data is crucial for a site’s ability to sell advertising, because advertisers demand detailed reports on who their messages are reaching.

While some journalists fear excessive attention to analytics will corrupt the quality of journalism, others believe smarter metrics can help news outlets connect with audiences in meaningful ways.

“The problem with statistics is there are so many of them, right?” said Audrey Cooper, managing editor of The San Francisco Chronicle. “Trying to find which ones are actually most useful and getting it down to a reporter level is not as easy as you would think.”

How then are media outlets judging the size of their readership?

  • “We want to start helping them [writers] realize how much of an impact they can have on their own readership using social media,” said Bill Adee, the vice president of digital stuff for The Chicago Tribune Media Group. “Any metrics that we can pull out that show that effect, that impact, is what we’re doing these days.”
  • “The No. 1 thing you want to know is: Are people actually reading this?” said Lauryn Bennett, head of brand at Chartbeat. “There have been a ton of studies that say the more people read, the higher likelihood there is of coming back [to the site]. If a visitor reads for three minutes instead of just one, their likelihood to return to your site doubles. That’s the holy grail.”

Many publications are outsourcing to large analytics companies, including Chartbeat, Omniture, Quantcast and comScore.

Bennett said Chartbeat alone works with 80 percent of top U.S. publishers, though that doesn’t mean those publishers aren’t also working with other companies as well. For instance, The Los Angeles Times gets reports from Chartbeat, Omniture, Google Analytics and Simply Measured, the last of which focuses solely on analyzing social media metrics.

Some media outlets are going even further and combining all their analytics into personalized consoles. Forbes Media and The Guardian were both early adopters of dashboards that fed their contributors and editors customized data.

Forbes’ “statistics engine” measures metrics, including pageviews, unique visitors and repeat visitors. These numbers are sent back to contributors, who have individual dashboards, as well as an aggregate dashboard for editors, according to an interview Forbes Media chief product officer Lewis DVorkin did with journalism.co.uk in December.

“The data informs what we do — it doesn’t rule what we do,” DVorkin told journalism.co.uk, “but not having it is being blind to what the audience is interested in.”

DVorkin did not respond to requests for an interview.


Related story: “Buzzfeed’s Secret Weapon: Ky Harlin”


The Guardian’s dashboard, Ophan, allows writers and editors to see updates much more quickly; before its development, the Guardian had to wait four hours for its analytics to update, according to a recent FastCompany article.

Ophan does everything from tracking traffic based on tweets to telling the site’s backend developers how long it’s taking to load a page on a phone or computer.

The Guardian’s digital audience editor, Chris Moran, said he believes in using data to inform the way journalists work.  He said Ophan helps reporters “determine the best channels for promotion” based on the content of their story and the audience they’re looking to attract.

The San Francisco Chronicle is on the same page as Moran. Cooper has mandated that all her reporters undergo a social media bootcamp, which began in early February. The paper is also in the process of creating individual analytics dashboards that show reporters simple metrics such as pageviews and unique visitors but also engagement time, social referrals and DMA (designated marketing area, which can show how stories are doing in a particular geographic area).

For some publications, though, looking at multiple standard metrics isn’t enough.

To keep up with the demand for a better way to measure audience engagement, Chartbeat’s new metric, called engaged time, tracks “people” rather than “events,” Bennett said, which is the preferred measure for Google Analytics and Omniture.

“We silently ping users on the site and can anonymously say, “Someone is reading right now, scrolling this far down the page, spending a minute and a half actually engaged on the page,’” Bennett said. “Are you scrolling? Are you commenting? Are you navigating? Are you actually engaged or have you just left that tab open and gone to get coffee?”

This precise feedback allows publications to know exactly which content their readers are engaging with — and hopefully gives them hints on how to bring that audience back.

Upworthy’s attention minutes, which measure “total attention on site” and “total attention per piece,” are trying to do the same thing.

In an email, Mintz wrote he was excited to see that Upworthy is not the only publication developing its own analytics tools. He hopes to see more outlets move in that direction.

Moran agreed.

“Data is there to help every journalist,” he said. “How wonderful to use it to maximize the number of readers exposed to challenging, important coverage.”

This story has updated comments from Lauryn Bennett, head of brand at Chartbeat.

Comments
  • http://bit.ly/MikeKing Mike King

    Great story. I wonder about (because this is something I’ve found myself doing) checking these metrics obsessively. I’m a writer for this website because of the class I’m in, and every time I happen to go to ajr.org, I check the shares for my article and they’re lower than those for other articles. It makes me feel like a social media loser.

    What I think might be being lost in the weeds with data-hounding is that sometimes really good stories exist and simply aren’t read by a lot of people for whatever reasons.

    I also don’t want to add more to every young journalist’s plate by saying they need to be checking analytics constantly. Journalism colleges tell us to write, record, video, photo, tweet, Facebook, Instagram, Vine, Google Glass and lithograph every single interview we ever have so we don’t fall behind the times. Are we now required on the front end to think about, “Well, if I write my story long or have more paragraph breaks, people will use the scroll bar more on their browser?”

    At what point does all this analytics hoo-ahh become so many dowsing rods, striking water occasionally but misleading through blind faith otherwise?

    • Lisa Rossi

      Great comment, Mike! Yes, we will be exploring these questions as we roll out more stories this week. I think there is a wide range of views out there on how influential analytics should be in terms of crafting a news agenda, which we will explore. Curious — what metric do you pay most attention to at the Diamondback?

      • http://bit.ly/MikeKing Mike King

        At The Diamondback, we’ve relied on some variant of page views for years, but as Mary Clare points out here, that’s not a perfect metric. We’re actually in the process of revamping not only how we work out metrics systems but also how we approach engagement overall, so a lot of these points are on my mind — hence clicking on this story when I saw it on Twitter.

        • Lisa Rossi

          Interesting… can’t wait to talk about this more… And thanks for clicking and commenting. Hope you scrolled too. :)

  • Palmer Brown

    Having been on both the sell-side and buy-side of website metrics the pageview has been on the injured reserve for a number of years now. If I receive an rfp asking for pageviews I assume (most often correctly) that I am dealing with a less experienced or unsophisticated buyer.

    Digital is not unlike print, outdoor, broadcast or any other media often referred to now as ‘traditional or old media.’ What matters most for both publishers and advertisers is reach and frequency. Pageviews can be part of the frequency formula but are meaningless otherwise.

    Buy-side view
    Reach (people): As a buyer I want to know what type of campaign reach I can achieve; targeted reach (demographics, geography, psychographics, buy cycle position) and reach waste (out-of-market: do not match the targeted dgpb). Unique visitors + other data = Reach. (-)Waste for determining efficiency.
    Frequency: How many times can I touch the targeted reach. This can be viewed over term or session. Visits/Unique Visitor and Visit recency/frequency is used to calculate campaign frequency. Pageviews are/can be used to calculate per session frequency. The most effective campaigns employ some sort of frequency cap.

    Sell-side view
    Whether you have publishing or yield management responsibility you should focus on building addressable audience segments for reach. Visit frequency, recency, and time spent should also be reportable KPI’s. Especially if you rely on parallel advertising for monetization.

    That’s all. I have work to do.