eMarketer, for example, estimates that more than 70% of US consumers now watch OTT content, although that figure is likely skewed by the inclusion of YouTube, which is best-known for short-form and user generated video. But it is clear that OTT is now a major force, particularly for on-demand or time-shifted content.
With OTT viewership on the rise, the next step for broadcasters and content distributors is to explore the commercial opportunities associated with the distribution of content to a growing diversity of connected devices. The monetising of OTT-delivered video falls into three primary buckets. The first is transactional: viewers pay a fee for each piece of content they view. The second is subscription-based: viewers pay a feeto accessa library of on-demand (and sometimes linear) content. The “all you can eat” subscription model is best represented by Netflix, but now also includes offering from an increasing number of content owners and distributors, which often include a mix of linear and on-demand programming.
The third route, and the one which appeals most to traditional players in the media industry, is advertising. This is because advertising has been extremely successful in funding broadcast, and it is easy to transfer those principles to new methods of delivery of the same content.
Advertising is becoming increasingly relevant to OTT content delivery, which is a personalised experience by nature. Consumers choose what they want to watch, when they want to watch it – and increasingly on a personal device. The personalized and one-to-one characteristic of OTT lends itself to a more targeted form of advertising, based on impressions, rather than ratings. Though advertisers selling ads in broadcast have been somewhat slow to adapt from a ratings-based approach to selling ads based on impressions, as OTT delivery allows, marketers are increasingly recognizing the value of reaching a targeted group of viewers with relevant information.
OTT is a one-to-one, unicast experience. In a well-ordered service, content providers should know something and quite possibly a lot – about each subscriber. Some data will come from registration, but there is the potential to acquire a lot more information from the subscriber’s behaviours. Leveraging the capability of big data analytics, media companies can now build up the profile of each individual, adding significantly more detail to the basicinformation collected at registration. The key to commercial success with an advertising-based unicast model is to understand viewers as individuals.
After acquiring that knowledge, service providers can target OTT viewers with relevant ads that are personalized to the individual, down to the device on which they are watching, be it smartphones, computers or connected TVs. A typical scenario would include sending an ad for a sleep-aid to someone at the airport consuming video on her mobile device before boarding an overnight flight.
That is why the media industry is now beginning to recognize dynamic ad insertion (DAI) as the holy grail of video monetisation. DAI, as the name suggests, inserts advertising at the time of delivery, so every single stream of a piece of content can be tailored to the tastes and interests of the individual. Technology to enable this is now in place, and demand is growing.
In conventional, linear television advertising, commercials are “targeted” at a broad spectrum of interests likely associated with the content. So a spot for a high-performance car might be shown during a break in a do-it-yourself programme, but probably not in a wildlife documentary. In addition to interests and demographics, advertising may also be influenced by geographies.
But this is targeting at the highest level. Advertisers pay on the basis of the number of broadly relevant viewers of their commercials. The fundamental metric is the cost per thousand (CPM), and broadcasters will attempt to drive up the CPM by demonstrating popular content which reaches the target market of the advertiser.
Targeted advertising using DAI turns this model on its head. The goal is not always to offer large numbers of viewers but rather high-value viewers who closely match the target demographic of the advertiser. Even more powerfully, studies show that consumers who see only commercials that are relevant to them pay a lot more attention to those commercials. Messages are much more clearly retained, making it more valuable to the advertiser.
Better targeting of ads is beneficial to all parties involved—advertisers, consumers, and content and service providers. DAI helps advertisers advertise more effectively, consumers receive ads they care more about, and service providers and content owners increase their overall revenue.
A key technological requirement for DAI to be acceptable to consumers is that viewers should not be aware that it is happening. Early models for DAI were based on downloading and storing replacement ads in advance on the target device, then switching from the stream to the commercial as required.
This results in freezes and buffering while the display switches from one source to the other, before the commercials and when returning to the programme. There may be a shift in perceived quality if the programme and the commercials use different codecs and bitrates. If the consumer feels that there is an impact on the enjoyment of the content, or that some of the content is lost in this re-buffering, there will be resistance. Software ad blockers will also see the switch and could simply eliminate the advertising altogether.
The success of DAI, then, depends upon delivering a consistent stream to the consumer. It also depends upon making the decision on which commercial to insert into an individual stream at the time of delivery: a spot which was relevant at the time the programme was originally made available may not be relevant during a future viewing, if say an event had passed or a marketing campaign had closed.
In order to replace a stale, general ad with a timely and targeted one in the stream, the location of the original ad must be known precisely. Without access to the detailed playlist from the original automation system, sophisticated discovery technology that recognises and finds all the ads in the video stream is required.
Once the potential ad-placement inventory in any piece of content is known, the DAI system can plan a campaign, and deliver advertising which is targeted to the individual, and which sits inside the OTT stream for a seamless viewing experience. For broadcasters that provide conventional channels and OTT services side-by-side, it is also important that the advertising sales and delivery process is part of a unified advertising management programme, allowing campaigns to be sold over linear and non-linear outputs and across delivery channels, targeted by demographics and by individual data analytics.
The technologies of dynamic advertising insertion are proven and continuing to develop. The ability to deliver targeted, relevant, fresh commercials in both linear and on-demand OTT environments is just the beginning. The future of DAI will be the unification of advertising management around multiple services and applications, including linear, on-demand, and cloud DVRs. The result will be coherent campaign planning and maximised revenues, however the content is delivered.