Personalized television
It goes without saying that personalized television is “in.” Subscribers increasingly demand content they selected, while advertisers are moving toward interactive, personalized advertising. The power of the Internet to customize ad delivery per user has set a new bar for advertising ROI. Few are feeling the impact of this change more acutely than the old guard of news and information, the newspapers. The Internet started as a text-based medium and now offers much of the same content found in the paper. Users have taken notice and are changing their behavior, with advertisers following their every move. As the Internet evolves from text to rich multimedia, savvy video service providers are adjusting their businesses to deliver the personalization demanded by the user while providing the audience segmentation and customization increasingly demanded by the advertiser.
Addressable advertising
Addressable programming and advertising represents a large opportunity for today's video service providers. While local advertising revenue continues to be a growing business for cable operators, it is being outpaced by the rapid growth for the advertising-supported Internet giants. Furthermore, the percentage of paid-click, or click-through, advertising continues to increase, representing a new category of high-value advertising not currently tapped by today's video service providers. Growth in ad spending has been slowly migrating from traditional broadcast outlets to interactive or personalized outlets. As the economy continues to evolve over the next several quarters, many anticipate that the pace of ad migration will accelerate. At the same time, the recipients of this shift in ad spending have announced plans to move in the direction of “display,” which is Internet-speak for television.
The demand for personalized advertising is an enormous opportunity for today's video service provider. Addressable advertising in the context of the home TV environment is the Holy Grail for today's ad buyers. Viewers are receptive and engaged, and personalized messages have high value. Video service providers therefore have a unique opportunity to dramatically increase the advertising CPM by providing advertising messages that deliver on that promise.
They can also tap the relatively new category of high-value paid-click or click-through TV advertising while providing the advertiser with unparalleled metrics about the viewing audience. Video service providers also have an important advantage over the Internet service providers when it comes to the subscriber. For Google to compete with the video service providers, it needs to make us change our lean-back TV behavior, our home network and our love affair with the remote control. Video service providers can extend their offerings to meet the 21st century demands of users and advertisers without asking much of the consumer. All they really need to change is the back-end technical infrastructure.
There are four high-level types of technology required to provide addressable advertising: the subscriber database, the personalization algorithm, the ad splicing technology and the required bandwidth. Service providers have excellent subscriber databases that support billing and marketing efforts, while personalization algorithms are readily available from a variety of vendors. This discussion will focus on the latter two types.
Ad splicing technology
Get the TV Tech Newsletter
The professional video industry's #1 source for news, trends and product and tech information. Sign up below.
Ad splicing, while relatively new to local operations teams, is a relatively mature technology. Standards for signaling ad avails and transferring files are well understood by a broad array of technology vendors. In short, splicing is easy. (See Figure 1.)
However, splicing ads into programs on a scale of granularity to match the number of televisions in the network is an entirely new challenge. Video processing and quality control must be separated from the ad splicing function such that video processors are allocated per asset (ad or program) and not per stream. Today's stream-based ad splicer is a fully equipped piece of video processing hardware, capable of changing the video pixels and macroblock quantization to create the correctly sized and shaped space for the inserted advertisement. While this stream-based approach has worked well for regional ad placement, it has two fundamental limitations for addressable advertising: quality and cost. By enabling these platforms to modify the video to insert the advertisement, operators lose control over the quality of the program and advertisement. The highly distributed architecture required to support personalized ad insertion makes the QC issue exponentially larger.
The costs of stream-based systems scale up with the number of streams. To enable ad splicing, every stream requires its own video processing unit. While this cost model works when shared across thousands of subscribers, it falls apart when each television requires a unique stream and a dedicated processor. To solve the quality and cost problems, there must be a fundamental shift from the legacy approach of stream-based video processing to a new approach of asset-based video processing. Ad or program assets can be preprocessed and stored, once and only once, as a constant-quality variable bit rate stream. They can then be combined by a scalable splicing multiplexer at the network edge in a cost-effective manner without compromising quality.
Required bandwidth
Addressable ads require addressable bandwidth, and bandwidth in a cable or satellite system, while extensible through incremental capital investment, is finite and in high demand. The introduction and expansion of HD programming is critical to the success of the enterprise, but HDTV demands four to five times the bandwidth of SD. HD VOD is a valuable revenue generator and competitive tool, but the bandwidth required to support it further reduces the available pool of bits. While it may be possible to store the advertisements locally on the associated set-top box/digital video recorder, the vast majority of deployed set-top boxes don't have the storage capability and could not participate. Therefore, to deploy an addressable advertising solution in the near term, the unique and personalized ads have to be delivered centrally, much like VOD assets. And while neighborhood segmentation will get us part of the way there, to match and exceed the value of Internet ad placement, video service providers need to provide customization at the household, and even TV room, level.
There are several bandwidth management tools available today. Satellite operators and some content providers have adopted MPEG-4 for HD programming. Cable operators are transitioning to digital and deploying switched video technology, but more is required. MPEG-4 is not practical for cable in the near term due to the large deployed base of MPEG-2 set-top boxes. Switched digital video will not work for satellite due to the inherent one-to-many design of the satellite transmission system.
It is clear that providing the personalization users and advertisers demand will require a holistic approach. For cable operators deploying switched digital video, much of the bandwidth gain is realized on the long-tail content, or content that is not highly viewed (not where advertisers typically want to buy time). So while switched video is a critical first step to addressable programming, the operator will need to first create the bandwidth/space for the new tier of switched services in addition to optimizing the switched programming for the most efficient bandwidth use.
First, video service providers need to take a careful look at how to optimize the broadcast spectrum, the largest part of spectrum use today. By applying advanced compression techniques, up to 50 percent of the broadcast spectrum can be reclaimed to support addressable programming and advertising.
Second, service providers must find methods to deploy the known and proven technology of VBR statistical multiplexing on the personalized programming that viewers and advertisers demand. By enabling VBR for these streams, another 50 percent gain of the switched/on-demand bandwidth can be realized and applied to addressable programming. The critical event is a shift from stream processing to asset processing and a separated, cost-effective edge multiplexing and splicing capability.
Finally, and perhaps most importantly, quality cannot be sacrificed. Today's premium consumers, those most coveted by advertisers, are increasingly aware of quality differences and are being aggressively marketed to on the basis of quality claims. Sacrificing quality to achieve addressability would be a strategic mistake, and the consequences would become quickly apparent.
Embracing a new approach
The opportunity to provide addressable programming and advertising is enormous. It is the lifeblood of the new media giants and a critical component of the 21st century media and entertainment landscape. The potential is great, and the enabling technology is within reach.
To provide an architecture and cost model that will enable addressable ad splicing, service providers must embrace a new approach to asset-based video processing. To provide the required bandwidth, they must optimize their systems holistically, squeezing every last bit from all segments of the spectrum. And to maintain a strong competitive position against up-and-coming service providers, they must continue to provide the highest-quality programming on the most important display — the television.
Chris Gordon is director of product management for Imagine Communications.