Going Beyond the Keyword with Interest Graph Topic Targeting

Business Blog / October 19, 2015

Topic Picker on phone

Topic Picker on phone

Today we’re excited to announce the enhancement of Flipboard’s ad targeting capabilities with the launch of our Interest Graph targeting engine.

Flipboard advertising has always been about delivering beautiful ads and great branded content in our contextual interest channels to maximize the impact of brand messaging to our users. We believe and have seen that the right ad (or story) in the right context for the right user can have tremendous impact for our advertisers.

With the launch of Topics in Flipboard 3.0, we developed great sophistication and prowess in understanding both the interests and passions of our users. We have a deep comprehension of how our users consume content about their interests and how interests are related to each other. These relationships now power our world-class content recommendations.

Today we take contextual ad targeting to yet another level by integrating our topic graph engine into where advertisements are served.

How This Works

A hypothetical packaged food company wants to advertise a Promoted Story and target it to the “Recipes” topic on Flipboard. Similar to targeting keywords on some other platforms. However, the similarities end there. With our graph targeting we would also dynamically serve ads to “Nutrition, “ Gardening”, “Grilling”, and “Baking”. Or even “Drinking” and “Breakfast”.


As we leverage the power of the Flipboard topic engine, our billions of reader, social, and engagement signals, we help marketers find additional contexts with relevance to their brand message.

Insights That Go Beyond Keywords

The example above and the relationships between Flipboard topics seem fairly obvious, but it’s an overly simplified example. The interest graph targeting allows our ad server to tap into much deeper insights.

Consider the topic, “coffee”. The interest graph identifies “espresso” as a related topic in which to serve ads. Obvious. However, as we explore the graph around “coffee”, we see that the topic “beer” is also in the local area. What is this relationship that our data has revealed to us? Coffee and beer are both beverages? It’s not that simple. Orange juice is not in the cluster. While we can’t dive deep into user psyches to understand precisely why this relationship emerges, we do have reasonable hypotheses. Coffee and beer both have devout followings of connoisseurs, drinkers of discriminate taste. There exists a craft roast and micro roast movement in coffee analogous to the craft brew movement in beer. The exact “why” to this relationship exists currently eludes us, but more important is that this relationship exists.

The users in this cluster can be a large group of people who follow both coffee and beer. People who follow and read only one of the two topics, but have interests in common content like precision brewing methods or fine brewing equipment. They could also have common interests in stories on Flipboard completely unrelated to beverages altogether.

Through the interest graph, we can help brands find and reach people in specific mindsets and passions. We can go beyond the keyword to find like-minded people who will resonate with an advertiser’s brand narrative.

Powered by Content

Flipboard is an amazing platform with 80 million monthly active users who are reading, sharing, curating, liking, and discussing stories each month across thousands of publishers. This gives us a rare opportunity to glimpse into the nature of content consumption and what our users really care about. With these new targeting features, we are opening the world of interest to our brand partners.

Dave H. is reading Olympic Weightlifting