Gannett’s USA TODAY NETWORK of 109 daily newspapers has 110 million unique digital visitors each month (40 million to the local papers and 70 million to USA Today), putting it in the exosphere of news and information sites. But these days Jason Jedlinski, VP of product management at Gannett and the Network, is more focused on quality than quantity — what those many millions of users want to read and why.
In this Q & A, Jedlinski details how he and Gannett, with a big help from technology, are working intensely to understand the choices that readers make when they go or are sent to an article, video or other piece of editorial content:
Gannett has a huge array of audiences on multiple local platforms. How do you sort them and connect with each user by attributes and behavior?
We continue to evolve our segmentation and targeting capabilities, which vary across platforms. We see it as an ongoing journey, especially as technology keeps improving. We’ve invested in research and data modeling to build a deep and actionable understanding of the benefits people seek from digital content, rather than just describing demographics and psychographics — or reacting to the last story someone clicked on.
We believe why you’re reading a story matters. Not only for what we recommend you read next, but how we present our content. For example, if you’re reading a story about healthcare reform because you want to plan your family’s budget for next year, you’re going to seek very different information than someone who is a political junkie, handicapping odds for the next election.
So, yes, in addition to the usual attributes (e.g., baseball fans, crossword players, video viewers, newsletter subscribers), we’ve built a well-defined understanding of what ultimately motivates someone to visit us. We’re leveraging those insights for our brand advertisers, suggesting effective content and creative strategies to make meaningful connections with our various audiences.
How do you make sure that your targeting algorithms don’t produce “garbage in – garbage out”?
We’ve done a great deal of primary research to develop our look-alike models. We validate our results and segments against third-party data attributes, to assess how well our view of users’ interests and attributes aligns with how our advertisers think about them.
We’re constantly evolving the ways in which we classify our content. Right now, we’re integrating blended sentiment scores to get a better grasp on tone and topic associations.
What’s so important about “attention time,” “hover rate” and “dwell time”?
We’re in the audience engagement business. Metrics like page views don’t necessarily correlate with what makes a valuable experience for a consumer or advertiser. We believe we can create premium, quality environments that are attractive to both groups.
From an analytics and predictive modeling standpoint, metrics like scroll depth help us inform the user experience. If most users only make it to the third paragraph of a story, that’s an important signal.
Have we satisfied a need? Could we be using a more compelling format? Are we interrupting the user and sending them elsewhere? Not all clicks are satisfied created equal, and we have to go beyond the simple tally of what headlines you clicked (or didn’t click) to understand how to deliver an experience that’s satisfying and worth coming back for.
Is the truism “Content is king” still valid — or should it be something like “A variety of content is king”? Or maybe “Video content is king” — and text content is perhaps a “princeling”?
I’d say we’re shifting to “Relevance is king.” Just as the ad-tech industry is focused on delivering the right message to the right user at the right time; we’re moving beyond curated, one-size-fits-all web pages to more tailored, personalized experiences that adapt to the user.
There will always be room for editorial judgment, but the way stories are presented can and should suit the reader. (Do you favor quick summaries and infographics, or a traditional narrative article? Do you often watch videos, or never click on them?) We believe it’s our duty to broaden horizons and stretch “bubbles,” but telling stories in a way that resonates is just good business.
Gannett pushes out a lot of USA Today Network content to its local daily newspapers. Is this done in part so the dailies can emphasize community-based content on their homepage, like with the Tennessean in Nashville?
Absolutely. Content sharing is a key strength of our “one newsroom” approach. We don’t need a dozen reporters writing about the Game of Thrones premiere. We typically feature community first on our local homepages and native apps, using social media to highlight our USA Today Network’s coverage of national and trending topics.
Signature franchises like “Humankind” highlight what are typically “local news” stories in our markets and make them accessible for a global audience. Our news websites share a common content management system, and we’ve focused recent upgrades on tools that help our reporters and editors quickly discover relevant content posted by distant colleagues. In fact, our Innovation Lab just built a feature that uses natural language processing to automatically recommend related content that writers may want to link to from their stories.
Is Gannett’s reorientation in user metrics producing quantitative results in monetizing content?
We’re still early in our journey, but yes, we’ve been able to increase revenue per visitor by more thoughtfully balancing the elements of a web page. In one of our homepage template tests, we found combinations of modules which added incremental placements for our advertisers while also improving user engagement and visit length. We weren’t expecting that.
Can this reorientation, over time, be part of a digital sub conversion strategy that leads to a new and important revenue stream?
Yes. We think so. We’re promoting the value of “membership” with our local brands, and we see the digital user experience as a strong component of that package, along with special events, deals and extras. We’re evaluating options like disabling auto-play video for subscribers, as an incentive for them to activate their accounts and log in.
Is Gannett doing anything new with Facebook and the other social and search platforms?
We’re taking the long view, treating off-platform channels as audience engagement and branding opportunities — not just immediate pageview generators. We know search and social present unique opportunities to connect with consumers who might not otherwise cross our path, and better-understand the touchpoints that prompt them to come back to us for more.
What unanticipated lessons have you learned that is helping you to develop best practices?
It’s always humbling to see how ingrained user behavior can become. We know people beat down paths around design hurdles. But even after you remove those obstacles, old habits die hard, and it’s surprisingly difficult to establish new patterns. That’s made us determined to avoid taking shortcuts we intend to “clean up” later.
We’ve also learned that we shouldn’t underestimate the appeal of something that had been hidden in obscurity. Poor performance is not always an indicator of future potential. Similarly, it takes more than prominent promotion to drive meaningful engagement. When pursuing actual relevance, versus classic KPIs like pageviews, relevance is king.
What’s beginning to emerge from over the horizon that will tell you even more about your readers and their behavior?
As an inherently local company, our ability to deeply understand local consumers’ behaviors is key to both serving their needs and forging meaningful connections with our marketing clients.
Slicing our audience data against variables like time and location has yielded a variety of actionable insights and product opportunities.
We’re running a series of experiments leveraging machine learning and ambient proximity technologies to help us get smarter about the connections and correlations of consumers’ habits, offline to online and vice versa.
We’re also bullish about the potential to extend our storytelling and local expertise into the physical world. That doesn’t require fancy augmented reality gear, but could be a podcast with a food critic’s tour of restaurant row.
We’re exploring a broad spectrum of premium experiences, which we believe can provide value to both our members and local business partners.
Finally, is the reorientation process giving you any feedback from your users that are useful in Gannett building better relationships with them?
Yes. We’ve invested in a robust, cross-functional “Customer Experience” program and we’re actively expanding the ways in which we collect and analyze user feedback, whether observed or volunteered.
It’s been particularly helpful as we’ve worked to improve common user journeys, such as signing up for a subscription on a smartphone, searching for an obituary, or trying to share a story you read in the morning paper. Both through their words and their behavior, your customers clearly tell you what you can do better. You just have to listen.
Tom Grubisich (@TomGrubisich) writes “The New News” column for Street Fight. He is editorial director of hyperlocal news network Local America, and is also working on a book about the history, present, and future of Charleston, S.C.