INdustry: Will Targeted Advertising on TV Be Better than on the Web?

For this edition of INdustry, we spoke with Eric Schmitt, EVP of Allant's cable and media business. Allant is a company to watch. An established player in the market services (database marketing, direct marketing, etc...) business, Allant has traditionally helped TV service providers manage customer lists and implement consumer direct marketing campaigns for old problems such as subscriber acquisition and churn management.

Now, Allant is extending those relationships and expanding its presence in the television business. They're applying their know-how in subscriber list development, segmentation and campaign implementation to create new resources and services for advanced advertising. Marketing HBO to cable households is pretty similar to marketing cars and credit cards to, well, households.

And this, interestingly enough, raises a major point.  

On the Web, video publishers can use behavioral targeting to make ad decisions. By comparison to the relatively blunt Nielsen model, behavioral targeting offers meaningful value-add to advertisers. A direct indicator of this difference: ads on broadband video garner significantly higher CPM's than broadcast TV for the same video content.

And this sets up the opportunity that Allant will help TV companies mine. You see, direct marketing, based upon actual household data, is significantly more accurate and granular than behavioral targeting.  

Eric gave an example: For most marketers, one of the most important distinctions in a group of people is "customer" vs. "prospect." In many cases, marketers have lists of their customers. So, given a list of people, it is relatively easy to separate the list into those that I know are customers, and everyone else.

Behavioral targeting, however, is based upon IP addresses. We don't really know who is surfing the Web. We know they went to Best Buy and Home Depot and checked out toaster ovens, but we don't know who they are and we don't really know why. Have they bought from us before? When? How much did they spend? 

The Web world can attempt to match the IP address to a person and there are many representations about the efficacy of this; but Eric asserts that, generally speaking, the match applications only succeed 20% of the time. (There is also the consumer creepout factor at the fact that their recent Web surfing behavior is clearly targeting their ad experience. This carries an attendant privacy/legislative risk.)

For a baseline comparison, a cable or satellite operator has name and address on file for 100% of its subscribers. The ID's of the set-top box in the house are all on the same file.  

The difference is clearly large for the advertiser that has his own lists, but it goes beyond the 20% vs. 100%. Household data can be leveraged extensively in the database marketing world. Marketing-services companies like Allant can append data from a large number of sources with a high degree of reliability. Palmolive may want retail shopping data. Ford may want income data, or likelihood-to-buy a pickup truck. Marketers differ widely in their strategies for targeting, and with household data, they have a huge number of accurate tools to draw from.

This is why media companies are now working with Allant to create viewer databases. Whereas previously, content publishers did not have a use for specific data about viewers, now the ability to connect marketers with viewers on a targeted basis is both viable and desirable. Perhaps best of all, the addressable model can be implemented consistently across TV, VOD and Internet video platforms.

Eric also provided an interesting observation about privacy. As he put it, when we deliver direct marketing pieces to the Post Office, we don't ask the Post Office for information about the recipient. We hand them the marketing piece with the address already on it.  

The "postal model" for advanced TV advertising highlights a key benefit of household-based advertising.

Marketers can run their own analyses, build their own campaign segments and so forth. Then they can simply instruct the TV network or service provider (depending on whose inventory they're buying) to deliver ad 1 to list 1, ad 2 to list 2 and so forth. While there are still many bricks to lay in such a system, household addressability makes that very, very easy.  

Behavioral targeting on broadband video has many virtues. It's already functioning at scale. It's lucrative. And it's growing rapidly. As the US TV market moves to implement TV Everywhere, we'll see broadband video flowing to homes authenticated by service providers. That will connect those viewing sessions to the operators' databases...and open the debate. Should be quite interesting to see how things unfold.



Michael is the president and CEO of MediaTech Strategies, a consulting company that helps investors and operating companies develop and capitalize on emerging media technologies.   He has over 18 years' experience in the media technology business, including previous CEO roles at Ucentric Systems, where he pioneered the multiroom DVR category and PhyFlex Networks, where he developed a new access network solution for cable.  Michael previously held senior executive positions with OpenTV, ICTV (now ActiveVideo Networks), Motorola and Playboy TV.  Through his consulting practice, he has provided services to more than 50 companies of all types, stages and size.

North America