A few weeks back, we looked at one of the biggest problems facing digital publishers—their ecosystem has morphed into one where traffic is increasingly being driven by clicks from social media sites, Facebook, in particular.
This means they lose control over the cadence of the editorial they put out since it’s up to the Mighty Facebook Algorithm as to when and where (and to whom) their stories are surfaced. (Facebook no longer shows links to everyone who “likes” a particular site—only about 20% of fans actually ever see a given link.) That means publishers lose access to data too, as they can’t track users around their sites and learn from their behavior.
Los Angeles-based IRIS.TV has been working to solve this problem for a while now, and their AI-based protocol should make publishers very happy.
The AI product is designed to take users on a journey through the publisher’s website, using machine learning to ascertain their preferences and then serve up videos they might be interested in. It goes beyond traditional tagging in that it knows that if a user is watching a video about the 2017 Super Bowl, they might also be interested in a music video starring Rob Gronkowski (“Gronk”), the New England Patriot’s colorful tight end.
By making these connections to create playlists that will keep users engaged and watching, IRIS.TV is hoping to create usage patterns that have viewers coming to the publisher’s website directly, rather than relying on Facebook as a launching point.
“The advantage to our AI solution,” notes IRIS.TV CEO Field Garthwaite, ”is that users who come directly to a publisher’s own website will then spend more time on that site. Much more time. That allows publishers to increase monetization. It also helps turn casual viewers into devoted fans, as they continue to see content they’re interested in and understand that the site is being personalized for them. That makes the experience much more valuable for both parties.”
The AI engine also helps publishers make better editorial decisions, as they can tell what type of video content users are drawn to, what shows up more often in playlists, which videos get viewed to completion, and which lead viewers to stick around to see what’s served up next.
We are big fans of what IRIS.TV is doing with AI and personalization. It offers publishers a way to take back control from Facebook and other social media sites while also creating better content. Rather than focus on what’s going to get Facebook users to click through, they can focus on the type of quality programming that results in loyal subscribers and a better environment for brands.
One of the key learnings from the programmatic revolution of the past few years is that context matters: the same user reacts differently to an ad depending on where they see it.
When journalists are free to be journalists and hew to an artistic vision, the stories they create are infinitely better and more relevant than ones they create solely for Facebook clicks. That’s the sort of content that brands want to run ads against because it’s memorable and meaningful and creates an emotional bond with the viewer—the type of emotional bond that also leads to better ad retention.
Using AI to drive recommendations is also a wise move. Too many recommendation engines rely on very basic tagging and thus miss important connections, serving up a generic playlist based on very broad factors: You like sports. You like fashion. You’re a Millennial.
With AI, IRIS.TV is able to get more nuance into its recommendations, surprising the user with connections most basic recommendation engines miss. And the more frequently the viewer comes back, the more IRIS.TV’s AI algorithm is able to refine its suggestions.
Publishers still have a long way to go to take back control from Facebook. But by focusing on quality (rather than clickbait) and by using IRIS.TV and similar tools to showcase and create superior content, they can start to to steer their own destiny once again.