This is presenting opportunities for brands, advertisers, content owners and broadcasters to derive revenues through more effective placement of advertising at times and places that fit with the overall viewing experience.

Cookies have been used from the early days of the public internet in the 1990s to track behaviour and target ads effectively on the basis of likely preferences. Meanwhile, broadcasters and pay TV operators denied such intimate knowledge of their consumers through having no return path at the time, resorted to all they had, context in terms of content being viewed at the time. Ads for running shoes might be shown during coverage of track athletics events.

Such contextual advertising was relatively crude and at first sight its re-emergence in an age when direct tracking of users had become possible did not create much excitement. But as privacy concerns have mounted, a seminal moment came in February 2020 when Google announced it would cease using third-party cookies in its Chrome browser before 2022, following the moves of other browsers in ditching this form of tracking technology. This sounded the death knell for use of third-party cookies in targeted advertising, and prompted development of alternative methods, with contextual advertising in the forefront.

A key point is that contextual advertising has become much more sophisticated than in earlier incarnations, exploiting analytics and machine learning algorithms to adapt to varying times, circumstances and content types. It has also gained from more advanced and granular analysis of content on the basis of key words in the audio and discrete objects or images inside video frames.

Many broadcasters are well aware that contextual advertising can deliver significant benefits and some have already conducted trials. UK broadcaster Channel 4 in 2018 conducted research under the heading of Contextual Moments, incorporating AI methods to place a brand’s ads next to relevant scenes in a linear TV show. This found that 62.4% of people exposed to the TV spot recalled the ad after prompting with screenshots if it had been previously been shown to them alongside an adjacent contextual moment. Only 31.1% recalled it after seeing the ad next to programming that was not contextually related in some way.

Now in the wake of the cookie bans, amid rapid growth in consumption of linear as well as on demand content from broadcasters on mobile devices, TV technology vendors have been seeking to exploit opportunities for more advanced contextual targeting there.

MediaKind, the video technology company created during 2018 from the divestment of Ericsson’s media business, has homed in on 5G as a major emerging opportunity for broadcasters, pay TV operators and content owners. The premise is that by delivering data faster and incorporating intelligent edge processing capabilities operating at very low latencies, 5G offers broadcasters, operators and D2C providers access to new data sources on their consumers’ connected devices that have yet to be tapped at all significantly.

MediaKind’s EMEA Director Market Development, Advertising Paul O’Donovan has been prominent in making the case for 5G based contextual advertising at various online seminars recently. The main argument is that mobile devices over 5G can make up for loss of cookies and add additional value by expanding content to include device location and other data gathered about a user’s immediate behaviour. The point here about 5G and edge compute is that together they can execute analytic algorithms and deliver results sufficiently quickly in near real time to enable on the fly contextual ad targeting decisions to be made. Then ads can be placed against content while it is still fresh in the consumer’s mind. One example might be to pop up with an ad for fine tea at the moment a pot is being poured on the screen, before the moment has passed.

In effect, contextual advertising over 5G gives the opportunity to create rolling profiles of users within a temporal window that reflects their likely attention spans and responses at a given time and place, even though their identity is not known. This can be more valuable than a static profile of a user aggregating their behaviour over all times, places and contexts. It is well known that a given person can respond quite differently to ads according to where they are, watch they are watching, and the time of day, as well as other factors such as season and prevailing weather.

MediaKind is one of a growing band of video technology vendors well placed to exploit this opportunity, having an established product for TV advertising. The company already provides server-side dynamic advertising insertion for Belgian IPTV service provider Proximus for example, supporting targeting on a per-user basis across live, time-shifted and on-demand content. Server-side advertising insertion appeals to many operators because it circumvents ad-blocking and avoids having to depend on client software in the user device. As such it provides a good basis for contextual advertising, so that the likes of MediaKind were in a good position to adapt their products for this emerging opportunity.