August 15, 2012 – Two recent advances in data applications for advanced advertising illustrate how the combination of ad targeting on the one hand and better selection of programming ad inventory for specific campaigns on the other are likely to work in tandem to revolutionize TV ad buying.
A key step in the direction of enabling use of viewer data to enable personally addressable advertising in on-demand and linear content was taken earlier this month by advanced advertising platform supplier BlackArrow. Its Software Information Service (SIS) module aggregates disparate sources of anonymous subscriber information to facilitate targeted advertising on any combination of standards-based vendor ad management, routing and placement platforms.
At the same time, FourthWall Media, a supplier of software supporting interactive TV and advertising apps in cable, has launched a new division devoted to collecting and analyzing data to help advertisers choose ad slots in programs that are more precisely matched to the audience they’re targeting. Here the goal is to discern exactly what live linear program or programs in what day parts in various types of markets are most likely to deliver the audiences advertisers are looking for.
The two developments serve as bookends to complementary trends in advanced advertising where data generated from set-tops, billing systems and other sources can be applied in whatever ways are best suited to evolving campaign strategies. While targeted advertising has been a big focus of the advanced advertising agenda, better selection of spots suited to reaching the right audiences without relying on dynamic targeted placement will likely move the industry toward higher advertising ROI faster than if targeting was the only way to overcome the drawbacks to traditional advertising modes.
Indeed, acknowledges BlackArrow CTO Joe Matarese, targeting of ads based on viewer profiles will begin with on-demand viewing and only later migrate to linear programming, probably starting with use in time-shifted “Catch-up” or “Look Back” applications before going into live linear TV. “We’re going out the door with on-demand as the first area of customer focus, and then we’ll be working back to live linear services,” Matarese says.
As previously reported (June, p. 1), dynamic placements in video-on-demand content have moved into the mainstream of MSO ad strategies, but such placements currently are made across all users based on the relevance of the ad to the anticipated audience for a particular program rather than on a per-user targeted basis. “This is the product that brings about the moment when we break out of the confines of strictly making ad decisions based on the program or on the geographic location of the viewer,” Matarese says in reference to BlackArrow’s SIS. “This is when you can construct campaigns for audience segments.”
The BlackArrow SIS is designed to work within operators’ multiplatform, multi-device TV environments, he notes. Utilizing interfaces supporting SCTE 130-6 and other standards it enables operators to normalize subscriber information and audience reports across multiple platforms and formats.
“SIS can interoperate with our ad router or a third-party equivalent,” he says. “It has been pre-integrated with our ad router, which is how we’ll be deploying with our first commercial rollout before the end of the year.”
The SIS is the caching center for all the data used in the complex targeted placement process to determine which ad should be positioned in a particular slot of a program viewed by a particular user. It enables targeting based on such viewer-, household- or device-related data as geography, demographics, marketing segmentation, membership or service level.
In addition, the product supports the capture and management of qualified audience lists that streamline sales, decision-making and execution processes. And the data stored in the SIS also becomes the basis for detailed reporting on audience performance and inventory supply.
In the targeted ad placement process the ad router plays the key role of directing communications among the various architectural components, including the SIS. When the ad router receives a placement request in conjunction with a viewer’s selection of a program, it contacts the Placement Opportunity Information Service to determine what all the ad possibilities are for that particular program and ad slot when viewed on the device the user is accessing. It gets the data from SIS specific to that user and then communicates the aggregated information to the ad decision system, which determines which ad should be played out from the ad server.
The lowest hanging fruit for getting this architecture up and running effectively is the operator’s in-house marketing program, Matarese notes. “Operators are definitely set up today to make use of SIS for their own marketing campaigns,” he says. “They can send promotions that are relevant rather than pitching somebody on something they already have.”
Targeting sales of their local avails on VOD programs is another obvious first step. But there’s also a national level of interest in use of targeting on VOD now that usage has reached mass scale, Matarese adds. “With volumes of on-demand viewing in the tens of millions and close to 100 million avails to fill, it’s meaningful to start segmenting audiences for targeting,” he says.
All of this will set the stage for moving the architecture into applications with linear content, he notes. “Once customers start applying these capabilities to on-demand content and then go from there into interactive applications [with on-demand content], it won’t be a big leap to look at time-shifted linear content and then to say, why don’t I plug the SIS solution onto the backend of my linear ad system,” he says.
This is especially true for IP-connected devices where targeted placements in linear programming will be easier to accomplish in the unicast streams used with IP versus the broadcast streams used with legacy pay TV distribution. But where traditional linear is concerned, the wide use of switched digital video platforms creates a unicast-like opportunity for placements in programming passing through the SDV system, Matarese notes. “Then you can work your way back through SDV to the full QAM-based environment,” he says.
Meanwhile, FourthWall sees an opportunity to greatly enhance ad-buying efficiency in the linear TV domain by bringing to bear data mining and analytical skills it has been honing for several years. The company, through its new MassiveData division, can provide ad agencies and programmers the information they need to improve advertising ROI by applying advanced analysis to the vast volume of data generated from set-tops tied to FourthWall’s Ad Widgets and TV Widgets businesses and from the customer database inherited with its acquisition of the Navic Networks ad management platform from Microsoft in May, says Bill Feininger, general manager of MassiveData and senior vice president of media measurement for FourthWall.
“MassiveData provides advertisers and programmers with an amazing array of tools based on the valuable datasets we have,” Feininger says. “They can look at historical performance as well as predict future relevance of programming for any given audience segment, and optimize their strategy based on this information. They can also validate and refine their strategy as MassiveData continuously updates these datasets.”
Through its various outlets FourthWall continuously collects census-level set-top box viewing data from millions of television viewer. MassiveData creates datasets from all this raw data by applying proprietary data-mining and analysis technologies, which allows the division to provide new metrics to advertisers and their agencies, TV networks and cable system operators that are distinct from the kind of data generated by audience measurement firms such as Nielsen and Rentrak, says Feininger, who was a senior executive at Nielsen prior to joining FourthWall two years ago.
“As we planned the MassiveData business we wanted something new and interesting that would exploit the rich vein of data we collect through our Ad Widgets and TV Widgets,” he says. “We derive over 1,000 attributes for each viewer for each set-top box and do it for every 30 minutes of programming to get a picture of what the audience composition looks like for each segment. And we can project that into the future as well. Now we’re refining those models and driving to that ultimate goal of providing great tools to give researchers or advertisers the picture of what the audience composition is going to look like down to micro-targeted attributes.”
As a leading provider of EBIF (Enhanced TV Binary Interchange Format) technology, FourthWall has long-standing relationships with various operating companies and vendors such as Comcast Media Center, Time Warner Cable, Charter Communications, Bright House Networks, DISH Network, Rovi and Motorola Mobility. The company supports interactive advertising apps through its Ad Widgets end-to-end advertising system and about a dozen TV apps such as caller ID and customer care through its TV Widgets, in the process amassing data from several hundred thousand set-tops. The Navic platform generates data from another 1.3 million set-tops, Feininger notes.
“We’re close to several MSOs working with them to collect data on a daily basis,” he says. “We can look at a particular market like Lexington, Ky. and say, here’s what’s happening on every set-top there for every time period and, using our algorithms, drive that out to similar markets to give our clients a detailed picture of what they can expect the audience characteristics to be for ad placements in any given program at any given time of day. We’re seeing great correlations between what our algorithms are generating for markets where we don’t have direct data and the proof data we purchase in those markets for testing purposes.”
In this fashion MassiveData can deliver rich detail about audiences for any given program in any type of market to help ad buyers determine where they’ll get the most bang for the buck, he says. “For example, let’s take an advertiser with a new product that’s done research and says the target market for that product has these precise attributes, which may include household income, ethnicity, age, location and other basic information,” he continues. “But they also may be looking for only the people with those characteristics that are theater goers or who are cycle enthusiasts or who have purchased a computer in the last thirty days. We can help them find that audience, which may turn out to be people watching a niche network during a particular time of day or tuning into an NBC program in prime time.”
As MassiveData brings in more people and refines its techniques it is talking to potential customers to begin the engagement process. “So far everyone we’ve talked to is very, very interested,” Feininger says. “Everyone has gone to the second stage to look at either getting right to the service or to do some kind of pilot.”