Turning Digital Dimes into Dollars: Online Video Advertising
Editor's Note: This is a vendor-written case study. StreamingMedia.com accepts vendor-written case studies based upon their usefulness to our readers.
Having worked in the streaming and digital media business for the better part of 15 years, I've seen and enjoyed being part of the tumultuous changes that the industry seems to thrive on. As with the climate ("If you don't like the weather now, just wait fifteen minutes!") it is mercurial; if you think the methodology we collectively use to deliver a particular feature today needs improvement, just wait a while. Technology may not change as quickly as the weather, but recently we've seen major advancements emerge within just months. This spectacular progress is largely the product of rapid deployments of TV Everywhere by both broadcasters and service providers.
In this new multi-screen world, we all want our media to be available to us anywhere, any time, and on the screen of our choosing. That's all well and good, but someone has to pay for it, right? Clearly, subscriptions help foot the bill, but the broadcast business has always relied on revenue earned from placing advertisements in and around content. Expect the digital realm to be no different: it's all about the ads.
Not Your Father's Television Set
We've all heard that "eyeballs are moving", which means the living room is no longer the central location for media consumption in the home. A majority of us now consume digital media on our smartphones, tablets, iOS devices, and the new generation of mini set-top boxes like Apple TV, Roku and Google TV. Make no mistake, content creators, broadcasters, and service providers know this and are developing their platforms and plans to meet you and your eyeballs -- wherever you plan to watch.
The new frontier of glass screens and pixels is clearly not the domain of your father's old television set; it's the "next new" with different models, features, and opportunities. Having said that, digital bread is still buttered by and large in the traditional mode, with packaged offerings targeting the living room. While it's hard to argue with business models that have worked for years, digital delivery of live linear content opens up new venues for advertising along with the tried and true.
For live linear content that is perishable, like news and sports, digital delivery offers the ability to insert ads in-stream, completely different from the ads used in terrestrial over-the-air broadcast. We need this functionality for a variety of reasons, licensing and performance rights being one, but also because the reach and cost models of broadcast versus streaming will change in the years ahead as eyeballs migrate and viewing patterns change.
What Is Old Is New Again
While we're delivering content and advertisements to new screens with new means, it's important to understand the technologies and processes within the existing broadcast infrastructure so we can discuss how these can be leveraged to help the new ad process.
For many years, broadcasters and service providers have used defined standards to signal the presence of coming ad breaks within television transmissions. These standards are researched and developed by a group called SCTE (the Society of Cable and Telecommunications Engineers). SCTE is comprised of personnel from many disciplines and fields of expertise in the cable and telecommunications industries.
In the cable and IP space, the signal that triggers ad breaks is called SCTE-35. These markers generally appear in the video bitstream roughly four seconds prior to the beginning of the break, and can contain a great deal of data about the ads, length of the break, and other pertinent metadata. In the broadcast realm, the signal is handled by a specification called SCTE-104. While the end result is the same, these two specs are somewhat different in their approach. SCTE-104 by definition covers a much broader spectrum of signals and means. In the simplest terms, SCTE-104 exists as a catch all for broadcasters to use a variety of signals, black boxes, red buttons on consoles, and switchers in order to enable ad insertion in linear and live television transmission.
One of the more interesting technologies to emerge recently leverages SCTE-35/104 markers to trigger ad insertion within the digital stream. By using existing workflows, the technology can be adopted quickly by broadcasters and service providers with a minimum of disruption.
Inside the (Digital) Sausage Factory
So let's look at the topology of how live ad insertion works in today's digital media architecture, including the role of SCTE-35/104 markers. The graphic below outlines the primary components in the creation of and delivery of streams in Apple's HTTP Live Streaming (HLS) specification.
The live video signal from the event or 24x7 channel is sent to a live encoding appliance, where it is processed, digitized, and formatted for the optimal audio and video resolution at various sizes and settings. These transport streams or "streams" are contiguous files that consist of H.264 video and AAC audio. We see these in all the prevalent formats.
The streams are routed through the appropriate servers for playlist manipulation and ad insertion to achieve the desired result. This approach is often referred to as "stream switching." In some cases, ad processes can also be handled by a network media processor (NMP), such as that offered by Envivio, that also performs other functions including digital rights management (DRM), packaging and encapsulation, or edge processing.
While I'm outlining the abilities of the HLS world, similar functionality is possible using Adobe's HTTP Dynamic Streaming (HDS) and Microsoft's HTTP Smooth Streaming (HSS). Although we are still in early testing and implementation with the MPEG-DASH standard, it is fair to make the assumption that the specification will provide for live ad insertion.
HLS, by structure, calls for some additional complexity when inserting ad units in this domain. The primary reason for this is the relationship between the content (where files are referred to as "chunks") and the playlist (.m3u8). Switching from the primary playlist and the inherent chunks to a variant playlist which references and provides the location for the ad components and then back again is a tricky proposition. We'll explore that more in depth in a moment.
Conversely, with the HDS and HSS technologies the ad units can be inserted at the client where playback happens, an approach referred to as "player switching." This could be on a mobile device, laptop, or set-top client, so long as they are provisioned for and capable of playback of those technologies. This player switching approach requires the playback device to have the power and ability to perform additional business logic device-side. In contrast, stream switching does not require the playback device to execute any business logic in order to play the in-stream ad. This isn't to suggest that one technology is simpler, more efficient, or easier to implement than another, but it is important to understand the subtle variations in how each is deployed and operated. In the case outlined above, the actual ad units have already been retrieved from the appropriate advertiser and network and are prepared, measured, and delivered via this infrastructure.
That means that when the SCTE-35/104 marker appears and is properly signaled to the player component, a call is then made to the appropriate ad server/network for instructions and where to look for the digital ad units. Depending on the implementation, these units could be pre-loaded or delivered in real time. Once the appropriate number of units is played, the player returns to the original broadcast by tuning to the original broadcast and associated streams.
They're Coming, so Build it Already
We may not be building the digital equivalent to Field of Dreams, but, as we've established, there is an emerging need to lay the groundwork, define the business cases and rules, and provide viable ad solutions.
The market for product and service-driven offerings around ad insertion and delivery is growing, and we've seen several household names deploy solutions to production in 2011. Greg Philpott is the founder and CEO of mDialog, one of the pioneers in the field of stream management for live and linear ad insertion. In fact, mDialog has one of the few technologies in production that leverages SCTE-35/104 tone-based workflows to enable ad delivery in real time, in a manner that is addressable and measurable. mDialog's first mover expertise with Apple technology has evolved from a core competency into a stream management business that now counts some of the biggest names in broadcasting as customers.
"I am always struck by two dynamics when there's a breakthrough in a new enabling technology (such as adaptive bitrate streaming)," notes Philpott. "First, amazing new capabilities are suddenly available for you to take advantage of, and second, the technology may be new but the core business model generally remains the same.
"The U.S. linear TV ad buy in 2012 will be in the range of $67 billion. Our job is to ensure the sell side can leverage this new technology to maximize the revenue yield of their premium content, and to ensure that brands and agencies (the buy side) can also take advantage of the new streaming technology to engage in a meaningful and measurable dialogue with the audience."
Philpott says most of his customers are setting goals to be as disruptive as they can, while remaining simple: taking the proven ad-supported broadcast TV business model and applying it to internet-connected devices. To achieve this and attract premium ad dollars, rights holders need to create live and on-demand viewing experiences that aren't TV-like, but are TV. It needs to be available where audiences are used to watching their favorite TV news programming.
Paving the Road Ahead
As I've outlined here, the road ahead is motivating, challenging, and (hopefully) profitable. While the fundamentals of ad insertion solutions are being defined, there are still a good number of voices to be heard, voices that will contribute input to these systems and standards. That is to say, as exciting as it is to look at what we're doing today and be proud, we still have a long way to go.
I think that the consensus in the halls and offices of broadcast technology professionals is that what we have today and what we are building for the near future is a great start. What remains for us as technologists, innovators, and digital media pioneers is to replicate those key broadcast ad methodologies and processes and improve upon them -- add some digital pixie dust, to use Disney parlance.
The digital world offers opportunities to take proven models and make them better, and to come up with new business models, as well. While it is hard to argue with ad models that have worked for over 50 years, it isn't hard to imagine the new levels of interactivity and commercial possibilities that streaming delivery can provide. Maybe Doc Brown from Back to the Future had it right: "Where we're going, we don't need roads" because the new roads will be the sky. We'll clearly need solutions, including for ads, but the possibilities are wide open.
Internal testing finds the software offers improved video transcoding time and quality.
As broadcasters look to provide TV Everywhere, Envivio shows that it speaks their language.