Satellite TV provider Dish Network is exploring more ways to match marketers’ first-party data with its own window into TV viewership following the launch of its programmatic exchange last fall.
Dish already offered access to 82 segments, like age, sex, income and a plethora of propensities, such as the likelihood of a household to watch horror flicks, to advertisers placing bids through its addressable platform.
Although the entire point of addressable TV is to target individual households, for the most part it’s been a manual process until now.
Dish wants to enable custom data matches for marketers who wish to activate their own first-party data, augment third-party data and target a buy within Dish’s OnPoint programmatic TV platform – not just through Dish’s direct sales team, said Adam Gaynor, VP of media sales and analytics for the company.
By the end of the year, it expects to create or license a data management platform to facilitate those exchanges.
Gaynor spoke with AdExchanger about Dish’s programmatic expansion and growing the total addressable TV pie.
AdExchanger: What was a key thing you learned through the programmatic pilot?
ADAM GAYNOR: We learned we needed more demand to help foster demand. When you have more brands bidding in your marketplace, it makes it a healthy ecosystem.
Now that we came out of the pilot, we needed to work with a lot more demand-side platforms (DSP), but if we had to connect our pipes to every single DSP, it would have been very time-consuming and costly. Lighting up BidSwitch (the bidder tool from Iponweb, Dish’s SSP) allowed us to connect our pipes to about 150 DSPs [beyond its initial three partners].
So by offering TV impressions within a bunch of different [DSPs], it allows a brand to connect all these different screens in a single buy.
Are you activating Sling TV inventory alongside addressable campaigns for more of an incremental buy?
That product is still growing and maturing. Everything we do from our standpoint on Sling is dynamically inserted, and not all networks enable dynamic insertion. We’re at the point where we’re looking at cultivating the data and making the buys that are on Sling more targeted and more addressable.
Ultimately we want to bridge the gap between both platforms, so if someone comes in looking for moms who drive minivans, we want to be able to find them and remain platform agnostic, whether they’re watching Sling, Dish or our Dish Anywhere app. Right now, our addressable product on the linear platform is sold separately from the opportunities we’re selling on Sling. That’s ultimately where we want to go, whether it’s over-the-top or video on demand (VOD).
Can you size up the addressable TV market today?
There’s the linear TV footprint – Dish is in 8 million-plus homes, AT&T/DirecTV’s at 12 million homes and Cablevision is around 3 [million] to 4 million homes. And Comcast does have some linear addressable business. In total, we’re probably looking at 30 million homes out of 100 million households.
But there are other platforms from the [cable and MVPD] companies, who are doing addressable on the VOD side, which is not through the same software we use. In total, there could ostensibly be 40 [million] to 50 million addressable homes at this point depending on how it’s executed.
Would you argue that the addressable TV marketplace is fully scaled?
When I look at where our revenue has grown, it’s a hockey stick. I definitely see addressable growing, but I’m not one who can or should predict the growth of the total marketplace. I can say my addressable business is scaling quickly.
In 2011, when Comcast was running some trials in Huntsville, Ala., there were about 15,000 homes being tested for addressable and everyone wanted to know when this would grow. Now we’re at 30 million homes in linear addressable? I’d say we’re at scale.
What does a typical addressable activation involve and what kinds of results are you pulling?
We did a campaign in automotive where the brand wanted to find people in-market for a pickup truck. Using Polk data, we were able to find in-market pickup truck intenders. On the back end, we were able to identify VINs (vehicle identification numbers) that were sold, match it against those households and create .07% lift using addressable. When you think about how many trucks were sold with that lift – over 1,000 trucks – and the [suggested retail price of the] pickup truck, it brought an incremental $26 million to that brand just from using addressable.
Does addressable ever not work for the advertiser in question?
Up until 2012, if you wanted to target someone with TV, you had to use programming as a proxy to get to that audience. When we first launched addressable, we had a handful of data points that said you can target against college education or a household with children. A lot of brands tried it, but then they wanted to bring their own data or use shopper data. Now they want advanced measurement.
In plenty of campaigns, the data comes back and shows nothing was sold or a measurement wasn’t positive. But that in and of itself is learning, and the brands who have worked with addressable since the beginning are still working with us.
What’s your biggest challenge with measurement?
We live in a world where multiple currencies operate against one another. I am for singular measurement for a brand that wants to measure across every screen. Dish has worked with Rentrak since early on, and we have a significant sized measurement panel of our households where about 4 [million] to 5 million households provide census-level, subscriber-level data.
ComScore has been processing that data for us since the start, meaning the currency of addressability is comScore. We can’t use Nielsen to monetize addressable, we have to use set-top box data. A few months ago we began licensing our data to Nielsen, which uses that set-top box data to strengthen its local DMA data. We subscribe to Nielsen so we can measure our linear advertising, but the set-top box data we give to Nielsen right now is one-directional versus Rentrak, which is two-way.
Interview edited for clarity and length.
