MICHAEL BOLOGNA: True addressable is defined as household level commercial messaging. Then there’s hyper-local targeting with commercials inserted down to the zip code or ad zone and then of course you have the mass national television. You have national TV to everyone, local TV to the DMA, hyperlocal TV to the ZIP codes that make up the DMAs and then you have directly down to the household. And that’s really the targeted television stratosphere.
Why do advertisers choose addressable targeting?
As a means to increase their frequency against a core customer target. Most advertisers have some type of mass television awareness and the reason they use addressability is to heavy up on their core customer without the burden of all the waste. TV advertisers typically buy based on broad demography, but we use addressability by taking data and matching it up against subscriber files and determine the segmentation of who’s living in the household. We use technology inside the boxes to insert the commercials only to those households we want.
What are biggest challenges?
If we only have 42 million households that have the capability and advertisers want to use this to reach very small segments, your segment (should be about) 25% (of that addressable base) or under to make this a relevant, valuable solution in most cases. So once you take 25% of 42 million, you’re down to (about 10 million). But that’s huge. Most addressable segments range between 8-17% so this is getting to your question about why the dollar value (of overall addressable ad market spend) is so low. If you’re only talking about a potential universe of 40 million and the subset of that, if you use it to maximum capacity is 3-5 million, on average, for a particular advertiser segment (and we’re only working with a limited amount of inventory – local two minutes per hour, which is owned and operated by those cable, satellite and telco systems) it really leaves a pretty small pool.
Even for advertisers using this technique as much as they can, their annual spends are still on the high end, south of $10 million. The high annual addressable spend for an advertiser is $8-$10 million because, these addressable households are still a subset of 142 million. You’re still reaching those households, but people use this as a means to hit the people they really want to hit without absorbing all that waste. At the end of the day, it’s challenging because you have to determine who is your real customer segment, then you have to match data against subscriber files and determine what percent of the universe is that segment but it’s got to be done across five cable systems who all use different technology, data sources and back end. To get that 5 million subset of the 42 million requires conversations and phone calls with 5-6 different systems all operating in different fashions.
I’ve heard there are huge disparities in pricing. How does an advertiser know if addressable advertising makes economic sense?
CPMs can range from $20-$500. Take a major credit card company, a high end credit card. They will buy national television and pay $25 cost per thousand for national television. What we do is take their customer files, match it against subscriber files to determine who’s not their customer. Then, of those who aren’t their customer, we determine which of them make enough money and have a high enough credit score to be their customer. When you figure out what percent of that population is, and create an effective CPM, that $25 cost per thousand is really about a $2,000 cost per thousand. There is waste in television and addressability doesn’t make sense for every advertiser. For an advertiser with a small segment, there ends up being a lot of waste, so a $75-$100 cost per thousand is actually very efficient because it’s a lot less than what they’re paying in national TV to reach that hyper-segment when you factor in waste. In many instances we determine addressability is not the right approach.
Can you give an example of when addressable is not the right choice?
If your target audience isn’t small enough and the cost of your product isn’t high enough, national television still is and always will be the most efficient way. If you want to reach everyone who washes their hair and your bottle of shampoo truly does cost $2.30, there’s not a whole lot of targeting that would make it more efficient than national TV because the smaller the segment the higher the price. In national TV, of course it’s more expensive to buy just women 25-54 than all adults 18-49. Same thing applies to addressable. It’s more expensive to buy women with children under 12 that make $60,000 than to buy women 25-54. It’s more expensive to buy households that are in the market to buy a Lexus, Mercedes or BMW than to buy adults 18-54. It’s using data and matching it against subscriber files to determine an audience of a very finite target and then using technology in the boxes to reach the target at the right frequency. Addressability is really about refining targets and minimizing waste.
Is there a difference between addressable and programmatic?
Most people who use the word programmatic haven’t the slightest idea of what it means. Combine that with every digital entrepreneur who’s made money arbitraging digital media now wants to do it in TV. We believe from GroupM’s perspective, the application and use of data to help optimize the way we work in television is fantastic and we support it. The automation of certain elements of the TV trading process is very important. The union of the two is what people refer to as programmatic. What we don’t support is auctions, bids, and what is happening in programmatic now is a lot of middlemen… and a lot of third party companies that want to buy television from the same people we’re buying inventory from.
Do you think there will be a time when all specialist or generalist DSPs hook in to linear TV inventory?
The guy in the middle wants to make money. Don’t get me wrong. There’s a need for people, data and technology in this exchange and a lot of those guys do it a hell of a lot better than agencies and advertisers. Our issue is the transparency and the arbitrage because television is very different than the Internet. Television still goes back to that spot and that unit in the show.
GroupM and WPP have invested a lot in TV data of late.
We use first party data, purchase-level data, third-party data, modeled data. We work with every data company from Experian to Dunnhumby to Polk to Acxiom to MasterCard. Whatever data source I need to help create and define the target segment that advertisers need, that’s who we work with. Having WPP’s involvement in companies like Kantar and Shopcom helps. Our recent partnership with Rentrak helps. It’s really just assembling a lot of different pieces to make this work.