AdExchanger: Where are you seeing the most growth in OTT?
STEPHEN STRONG: Apple TV, Pluto TV and Roku are our three biggest drivers right now, as well as Sling TV, and we’re aggressively launching new partners all the time. Our Comcast deal is two-pronged. Watchable was the initial catalyst, and we’re also one of a few partners for the MY XFINITY portal.
Our focus with Watchable was increasing share of voice and interactions with the editorial team, but one challenge when comparing new platforms to old platforms is that the metrics are different.
How does that impact measurement?
We work with a number of platforms, and the metrics we get from Vidible – now AOL – on desktop are different than the report we get from Roku with Sling TV at the end of the month. So, a lot of that data aggregation has been manual. But something we’re looking to do across the board is integrate our first DMP on the back end, where we onboard third-party data and match it on the back end with our first-party data.
A sub-point to that, in OTT in particular, is: How do you aggregate data between partners? As we look to maximize yield on OTT and folks shift dollars there, a lot of that is reliant on having a deep set of data. It’s a really big priority for us, especially when we have RFPs that come in, saying, “I want to buy Roku within eight miles of this ZIP code.” Having the data to support that presents significant opportunity in terms of CPM fill and driving more value.
Who are you using for programmatic video and OTT ad serving?
When we relaunched our site, we used Ooyala, which is powering the experience on Newsy at the moment. A lot of our syndication business is done through Vidible/AOL. There was concerted effort early on where we were [mostly] programmatic, which has its advantages and disadvantages. Tremor was one of our key partners, and they still play in the mix. We’ve done a lot of work with SpotX and BrightRoll. It’s about keeping those relationships to ensure we have the right mix.
Are you creating more private marketplaces?
Growing scale and constantly improving quality presents PMP opportunities. The good thing about Tremor and SpotX is that we have a lot of transparency into who’s buying our inventory. We can go in and see that one month, an advertiser spent $70,000 with us through direct deal ID or the public marketplace.
But how do we go in and take these premium advertisers that have a real interest in the Newsy brand, put them in a PMP and increase value for both sides? That’s something we’re starting to dig in to now, but I think there’s huge upside there.
How you prioritize what video gets syndicated and what’s socially distributed on a platform like Facebook?
A few years ago, like any content producer, you would create a video that’s a minute or a minute and a half and distribute it across all platforms. It wasn’t the best strategy for us. We built out a social team within our editorial team, so that what we’re publishing now on desktop is a little bit different than what we’re publishing on Facebook.
We have 55 folks who are true editorial, and we have a VP who runs our brand studio. We also just hired a producer who is hiring for that team. We keep that divide of church and state because we wouldn’t want the content we’re producing on the editorial side to be influenced by what we do in revenue.
On social, we’re doing a lot more video with text [overlays], so if your sound happens to be on, your video is now narrated. Social is a great audience play now, but it’s not a huge revenue driver, as most folks would say. But we’re building out a strong foundation for when revenue comes to Twitter and Facebook video in a big way.
Which data sets are you using to help validate cross-screen audiences?
Although we’re an OTT-first service, in the near future, we plan to launch the service into cable platforms, which is a big win and goes back to the data piece. The low-hanging fruit is plugging in to someone like WideOrbit to do the trafficking, or someone like Nielsen. But the challenge with Nielsen is: If you launch in smaller markets, are they accurately represented? And then you have comScore and Rentrak, but Rentrak isn’t in all markets.
It’s challenging, especially when you start out small and don’t have the national cable distribution, which balances some of these concerns out. You also have other (set-top box data) providers, like FourthWall.
But when you talk to a traditional agency and buyer [about trying a] a different data set and pairing it with something that may even be more accurate or granular than the big providers, there’s a sort of resistance. Unfortunately, I think getting the whole industry to [evolve] is a slow process.