Understanding the right content mix for digital consumers is crucial to any publisher’s success. Knowing when consumers want to consume video content versus text content, and understanding how different devices can determine how audiences move throughout your content ecosystem can be game-changers. Outbrain COO David Sasson spoke with Beet.TV about these topics and much more. Click on the video above, or skip to the transcript of the conversation if you prefer, to hear David’s perspective on these matters.
Our video offering has been adopted by many of the large publishers we work with from CNN Money to Fox to ABC to Reuters. By working often within their player environment and on their video sections we’re seeing all of the behaviors of which sections of their audiences really want to watch videos and which segments of their audiences want to stay in a text, print world. So we’re able to differentiate between those populations to a degree and make sure that when people come back to the site we give them the right mix of content types as part of our recommendation package. So for us, the video offering is a hugely important offering and it is part of a wider ecosystem of content recommendations and knowing how to get the right piece of content to each person in the format that is most appealing to them.
We don’t use straight demographics in terms of registration data in terms of whether you’re a man or a woman, age 10 or age 30. What we try and do is we try and see how much data we do have about you personally: if you’re a repeat visitor to a publisher site, and you come there over and over and over again every day we have a lot of data about the subject matters that you’re interested in, the type of content that you like to go into, and we can use that information to find highly personalized links for you. For a lot of visitors to a publisher site, as you know, they don’t come in over and over and over again. They come in once a month, twice a month, three times a month, so for that segment of the population what you have to use is a good targeting mechanism, not information about them but rather try to string the limited information you do have about them together with information about wider groups of people who’ve experience similar content offerings as them and make bigger inferences. In that sense it’s a little bit like — you hear about the social graph on Facebook and Twitter, who are the people you’re connected to, who are your friends — we take a similar methodology to that, except that these are not your friends, nobody that you know, but rather people that tend to consume the same types of content offerings that you do. We plot you on that graph and make some inferences about where you may want to go next. We take a very data-driven approach to algorithmically figuring out what to target to you but the amount of personalization depends on the amount of data.
Our strategy is to a large extent the same as a publisher’s strategy because we really are a toolkit for them. We have a mobile offering. We invested in that over a year and a half ago, two years ago. We power content recommendations on the mobile optimized screen, on smartphones, for many of those same publishers I mentioned to you beforehand, the CNNs and Foxs of the world, where when you get to the bottom of a small screen article on your mobile device there the problem of navigation, of finding something next to read, is even harder than on the Web. You don’t have a left rail, you don’t have a right rail, it’s not easy to back up into the home page and go back through. So by us placing a few very targeted links to what you may want to consume next at the bottom of that experience, we’re seeing very, very high engagement on mobile. We separate mobile out into different realms, the smartphone realm, the tablet realm, tablet is really not the same experience at all, it’s very different. We have offerings that have been used and are growing very quickly in those areas.
I think one of the underrated aspects of sharing that publishers are starting to pay more and more attention to is the actual titling of the piece of content. People usually share content because it’s good, because it’s meaningful, because it’s interesting, because it’s new. But that’s only half the equation. The other half is once they share it, how do you make sure the audiences who see that shared link will click on it, and will actually come back through and consume that piece of content. The titling that you use to represent your piece of content, and the thumbnail to some degree, making sure you have a thumbnail above that piece of content that will get pushed into a Facebook environment or into other areas, using pieces of creative assets like that effectively is more and more important for making sure you get the bang of the buck for sharing. We look at click-through rate on a site, we present links of content, recommended links, and we look at how many of those, what percentage of those get clicked. One of the things that have been ignored for a little bit is how does that kind of operation happen in a social environment. It’s not enough to just get it shared. That’s the first step but the second is how do you make sure it’s interesting and enticing enough that people will click back to it once it has been shared.
Different Devices for Different Audiences
Different audiences and different people want different types of content at different times of the day on different devices. It probably goes without saying. A lot of people probably intuitively think that way. People who are reading long-form articles about business during the workday generally don’t want to click on a video link. People who are leaning back on an iPad at 8PM at night reading about lifestyle often do want to consume a piece of video and don’t want to read a long article. So what we’ve been doing algorithmically is looking at those types of differences per person, per time of day, per device, and figuring out when somebody comes to the end of a piece of content should we recommend to them a video, should it be an article, should it be a photo gallery, what type of content is important to them at that point in time, and mixing all of those approaches together to get the right type of content out to users to satisfy their curiosity at the right point.