Long reads


Ad and divide

The use of two-way networks to deliver targeted advertising remains more a promise than a reality, despite the impact of the internet and the economic downturn on TV ad revenues. Stuart Thomson investigates.

The TV advertising business is not having the best of times. It is seen as having been outpaced by the more nimble internet, which has captured a growing slice of overall ad spend even as the current global recession has deepened. Financial results from publicly-listed commercial broadcasters have made depressing reading in recent quarters, as revenues from ad sales have tumbled

And yet TV still occupies a greater proportion of leisure hours and reaches a greater number of people than the internet. It remains a powerful medium by which to communicate marketing messages. It might therefore seem reasonable to suppose that advertisers would jump on the chance to use TV to target particular consumers that may be interested in their wares – something that is theoretically and practically possible in the case of addressable two-way networks.
Yet movement from small-scale trials of targeted advertising to large-scale deployments remains elusive. “Targeted advertising is in a chicken-and-egg situation, as currently there remain too few addressable viewers to make investment appear to be a worthwhile,” says David Price, vice-president of business development and marketing communications at Harmonic, which provides a range of technologies, including the Streamliner video and ad serving platform and ProStream 1000 ad splicing technology, that support targeted advertising. “Until either the advertising industry begins to demand targeting or sufficient unicast subscribers are available to target, this situation is unlikely to change. From a regulatory perspective, privacy is still the greatest cause for concern. Voluntary registration of preferences and interest will circumvent this so that ‘opt-in’ ad-supported services will be prevalent at least initially.”

Linear advertising

The US is by far the largest TV advertising market. Every hour of programming on US TV includes 16 minutes of advertising, of which 14 minutes are controlled by the broadcasters with the remaining two minutes controlled by cable operators. Adverts are inserted into the programme stream at cable headend sites. This has created an annual US$6bn (44.2bn) business for US cable – a business it is keen to defend and expand. However, the development of targeted linear advertising in particular – the largest potential market – has been slow to take off. “Delivery of addressable adverts to smart phones and other connected devices is currently driving growth in the industry with ads being linked to news, sports and other on-demand assets,” says Price. “Linear channels are always likely to prove more challenging as the infrastructure is set up to deliver content en-masse and to change this would require significant investment at both the head-end and last mile.”
Others are more upbeat. Technology company Arris has been involved in the digital programme insertion business for the last eight years. “It’s been a very healthy business for us – we have a 70% market share,” says Paul Delzio, director of on-demand product strategies at Arris. Advertisers and operators are now looking at how to use the linear platform to deliver more targeted advertising, according to Delzio. Basic targeting can involve the delivery of local ads after a national ad, for example. However, says Delzio, advertisers are increasingly looking at replacing an entire ad with one that is targeted to a specific geographic or demographic subset of their overall target market, allowing advertisers to tailor their message according to the income level of the viewer, for example.
For widespread adoption to occur, operators will need to be confident that their infrastructure investment is going to pay off. According to Harmonic’s Price, standards will be crucial in making the business viable. “For targeted advertising to become viable, there has to be a single set of standards to which everyone adheres. Breaking down addressable audiences into fragmented catchments each using differing standards will prevent advertisers in particular from engaging as they need to see a single unified and aggregated group to which they can individually target adverts,” he says. “SCTE-130 is the standard being adopted by US cable operators and it is a carefully structured set of standards that has the potential to expand to be embraced by operators of all kinds – not just cable – around the world – not just the US. A simple extension of this standard to add ‘location information’ to the data about a user allows for it to extend to mobile applications. The greater the adoption of one prevailing standard across all platforms, the faster targeted advertising will come to a screen near you.” Nevertheless, many products currently involve a proprietary element. Yossi Weihs, director of advertising products at interactive TV technology specialist SeaChange, points out that SCTE-130 is still evolving, and no-one has yet done a real-world rollout based on it. He says the company is currently working with one customer on integrating its VOD ad sales system with a third-party campaign management product using SCTE-130, but adds that this is currently a very expensive custom project. Weihs also points out that SCTE-130 has very little to say about how to deliver an ad into a VOD platform.
A number of other technical hurdles could slow take up. Bob Wall, product manager for advanced advertising products at Tandberg Television, whose AdPoint software is designed to deliver advertising into VOD platforms, says that a requirement to put new infrastructure in place may disincentivise broadcasters and operators. It could be possible to use existing business systems, but there is a certain amount of inertia. “The systems and infrastructure are different from the linear world. Traffic and billing systems could be extended to support these new areas but it has been a challenge to get those suppliers to improve their capabilities,” he says. Service providers too can be reluctant to take the plunge because they don’t know when the money will start to flow in.

VOD advertising

The investment necessary to introduce wide-scale targeting to the linear platform, and the rapid growth in take-up of video-on-demand services (driven especially by the availability of ‘free’ on-demand content and the introduction of start-over and catch-up services) has led US cable operators in particular to take a more immediate interest in introducing advertising (targeted or otherwise) to their VOD platforms.

“Targeted advertising is in a chicken-and-egg situation, as currently there remain too few addressable viewers to make investment appear to be a worthwhile.”
David Price, Harmonic

The rationale is straightfoward. ‘Free’ VOD, including catch-up TV, is increasing in popularity. The introduction of dynamic ad insertion on the VOD platform would enable operators to create wholly new advertising inventory. In the US, about 38 million set-tops that can be used to view VOD are in circulation.
“The number of set-tops in homes is growing by close to 30-40% a year,” says Delzio. “VOD advertising is the most exciting part because it’s creating inventory that did not exist before.” The inventory of programming available on VOD is growing rapidly, and the number of VOD streams requested is growing even more rapidly. “If you combine all these forces and you have a technology that allows dynamic ad insertion either pre-roll or post-roll you have suddenly created a large inventory of ad spots,” says Delzio. This advertising can also be targeted. Every VOD request is identifiable at the household or even individual level, and advertisers have access to a lot of public information about individual householders that can be integrated with the ad decision server. “VOD promises you targeted ad placements that are measurable,” says Delzio.
Tandberg’s Wall says that the market for VOD advertising, though relatively small overall, is nevertheless attractive. “If you look at the numbers, VOD is huge compared to where it started but relatively small scale compared to linear broadcast,” he says. “The interesting thing is that much more is known about subscribers that are choosing to use on-demand services. And it allows a way to monetise that content.”
The introduction of targeted advertising to VOD is not restricted to the US. In the UK, cable operator Virgin Media recently conducted a trial of advertising on its VOD platform to 100,000 subscribers, using another 100,000 that did not receive ads as a control group. One of the key results to emerge, according to SeaChange’s Weihs, which supplied the technical infrastructure for the trial, was that the inclusion of advertising did not result in reduced usage or increased churn. In addition, he says, only a small proportion fast-forwarded through the ads. The ability to measure and monitor the way in which users interact with the content is particularly useful to advertisers, as the number of VOD users (while small in overall terms) is larger than panels used by audience measurement agencies such as BARB in the UK, he says.
Advertisers are certainly likely to be interested in how viewers interact with ads. “If people start skipping over the ads that can tell you it’s time to refresh the creative,” says Weihs. “It’s very expensive to get that [information] out of the linear platform but it’s free on the VOD platform.” Virgin Media is currently digesting data and distilling lessons from the trial. Questions still remain over how many ads does it make sense to place in a VOD stream (and where to place them) and – crucially – whether the platform will make commercial sense in terms of traditional advertising metrics such as cost per million.

Degree of targeting

While VOD advertising in particular could be used to provide information back to the advertiser, the value of being able to target specific user groups (or even individual users) with particular campaigns is difficult to gauge. So far, little of this has actually been done, even in the US. ‘Targeting’ in the context of VOD advertising has typically meant placing ads in content that is the best fit for the type of product being advertised, or possibly in placing ads for a particular product in a VOD stream that’s requested at a particular time of the day  – in other words exactly the same thing that happens in the traditional linear broadcasting world where people buy slots in particular programmes broadcast at particular times of the day. Arris’s Delzio admits that targeting to date has largely been limited to the type of programming or the physical geography of a particular market (for example urban versus rural areas). However, he says that the integration of third-party subscriber data into the ad decision process is not far off. He argues that the drift of advertising away from TV towards the web will force the industry to act.
Gideon Gilboa, product marketing manager for advertising solutions at interactive TV technology provider NDS, believes that targeting will prove to be attractive. “The fact that you can reduce the amount of irrelevant views in a particular profile is already an improvement on what was there before. If you have an ad for nappies, then only households with kids will see this ad.” For large companies such as Proctor & Gamble or Unilever, the next phase could be to deliver multiple ads in the same slot to different target groups. Companies could also segment their market based on income level.

However, according to SeaChange’s Weihs, the idea that an internet model of advertising, where ads could be placed according the viewing habits or demographic characteristics of individual viewers, is misplaced, because it simply doesn’t reflect the way advertising on the TV works. Advertisers don’t know exactly which part of the audience an ad will resonate with. It is therefore likely that if they have the option to buy only a portion of the available audience (for example, high-earners), the premium they will pay for that degree of targeting will not compensate the broadcaster or service provider for the loss of a sale targeting the remainder of the subscriber base. “It’s a zero sum game at best,” says Weihs. “You could sell 20% [of your subscriber base] at 100% mark-up, but the rest will never get sold, and you will have deployed a lot of expensive technology and not made any money.”
It is likely that the existence of the internet and the competition it provides for advertising spend will nevertheless force TV companies to innovate in some way. Tools do exist, notably in the US, to enable agencies to better ‘target’ their advertising by providing detailed socio-demographic data about the audience of particular shows in particular regions. NBC Universal recently signed up to use Microsoft’s Admira ‘TV marketplace’, which  uses a database of demographic information about the audience of TV shows to tell advertisers what to buy and when if they want to target a specific demographic group. Admira provides automated planning, buying, posting and billing across NBC’s national broadcast and some local TV properties.

This kind of approach makes sense, says Weihs, because with the fragmentation of the audience and the proliferation of niche channels, targeting based on matching content to its likely viewership is becoming necessarily more sophisticated in its own right. Trying to do more could be counterproductive. After a certain point, he says, “you get to the point of diminishing returns because targeting is already happening at the linear level”.
SeaChange has a partnership with Visible World, which provides software that enables advertisers to deliver different campaign messages to different audience segments, down to the household level.

Business models

The major US cable operators – Comcast, Cox and Time-Warner – have begun trials of targeted advertising on their respective VOD platforms. Cox is placing ads in its MyPrimetime service, although it has yet to unveil results. Comcast has experimented with addressable advertising in both the linear and VOD tiers. However, it has in fact stepped back from targeted linear advertising, while progress with targeted advertising in VOD remains slow. Time Warner Cable is meanwhile developing dynamic ad insertion on the VOD tier as a product to offer to advertisers.
Aside from technology, a greater problem is that no clear business model exists. Cable operators can only sell targeted advertising in their own limited ad sales windows, or they can create new inventory around VOD assets.
Harmonic’s Price believes that an operator-centric localised model will be the place to start. “A localised business model approach is most likely to emerge around targeting adverts. This means there will always be a space for operators to insert and monetise ads on a local level,” he says. “At the same time, this should help grow revenues as they can increase the value and thus the price of advertising. This style of local ad selling already takes place with broadcast networks selling ‘avails’ that can be localised, within a national stream. Unicast IP-delivered video allows for any level of localisation and thus potential expansion of the value proposition.”
An alternative to cable selling its own avails is to provide a platform service and leave the actual ad sales business to the content provider. “Are the operators the ones with the inventory or do they charge [the networks] a fee? That part is not clear,” says Wall. “There isn’t a single model any more.”

[icitspot id=”9943″ template=”box-story”]

According to Weihs, advertisers will be willing to pay for ad inventory in VOD assets or in content held on DVRs if the price is right. However, it is the service provider that will need to make the investment to enable it. “Given the current economic climate it’s harder to make that decision,” he says. And most importantly of all, it will be up to the channel suppliers, who historically have sold the bulk of advertising inventory, to integrate their own business systems with those of the operator and make sure that information about the breaks is communicated to the VOD system. According to Weihs, US networks have been reluctant to commit so far because they are afraid that the technology platforms are not mature enough and they do not want to sink investment into platforms that will need to be replaced later. The major problem is with campaign management and business systems rather than the insertion technology itself.
Arris’s Delzio maintains that the broadcasters have been slower than advertisers to see the merits of targeting. “Advertisers are on board with anything that helps them reach their audience with a product or service – they want detailed measurement and accountability,” he says. Broadcasters, he says, are used to the existing business model, which in the US revolves around the annual sale of ‘upfronts’. Nevertheless, he says, this model will come under increasing pressure and broadcasters will also be forced to act. This view is echoed by NDS’s Gilboa, who argues that the industry will be forced to change, even if the business case for investing in interactive platforms is not yet cut and dried. “The model still isn’t clear,” he admits. “We can only guess how it will end up but there are drivers that are changing things.”

The US cable industry’s attempts to find common ground to market targeted advertising have met with mixed success. Much attention has been given to Project Canoe, which initially at least was seen as the foundation of a national interconnect network that would allow operators to deliver targeted VOD advertising across the US, allowing them to boost the existing US$6bn cable ad sales business. (In the US a single metropolitan area can be served by a number of different cable operators). Delzio remains confident that “by 2011-12 there will be a national advertising platform for the cable industry”.
SeaChange’s Weihs is however sceptical about Canoe’s prospects in this area, and believes that most of the inventory will be sold by broadcasters. “One thing about VOD advertising was that initially systems were put in place in the expectation that cable companies would sell ad inventory in VOD,” he says. “Now it looks like the bulk of the inventory will be sold by content providers.” He says that it is more likely that the cable companies will provide platform services to broadcasters and leave the marketing of ad space to the latter.

IPTV operators

If cable still has some way to go before targeted advertising becomes a part of the everyday experience of subscribers, perhaps IPTV providers will galvanise the MSOs. IPTV operators enjoy the benefit of a more technically flexible and sophisticated platform (though as yet without the scale that would make targeted advertising worthwhile). John Johnson, founder director of UK targeted advertising technology start-up The Zap Corporation says that the use of subscriber data by IPTV providers could lead to “a shake up of the value chain that will see network operators earn more directly from advertisers”. Zap uses Time To Live (TTL) technology – a protocol designed to keep networks free of clutter caused by random packets that get ‘lost’ in transit – to insert specific values into the packet headers that enables this protocol to be used as a targeting mechanism.
The model being promoted by Zap involves customers opting in, allowing their data being used to deliver targeted ads in exchange for some benefit in kind. “The toughest thing is the data privacy piece,” says Johnson. He believes the industry is still about two years away from developing a commercially viable business around targeting. For Johnson, targeted advertising is a way that IPTV providers could differentiate their offers from cable and satellite competitors. “Most people still struggle to figure out why there is a need for a fourth platform – there is a need for a couple of messages that the consumer finds easy to understand,” he says. “Advertising tends to be where the answer lies when the question is how are you going to monetise this?”
Bob Lamb, chief technology officer at ad scheduling systems specialist Pilat Media, agrees that IPTV operators are better placed than cable to deliver and monetise their platforms in this way. While IPTV providers can target to the level of the individual set-top box, cable networks typically are capable of segmenting their subscriber base only in a limited way, he says. “I know for a fact that IPTV vendors can go down to the set-top itself; cable operators may not be able to target down to that level,” says Lamb. “For IPTV operators it was simple because they built their infrastructure knowing they would want to go down to the set-top level. If you have an HFC network it’s much more difficult.” In the US, regulatory restrictions mean that IPTV providers can only target down to the household level (rather than the set-top). Nevertheless, their ability to target in such a granular way could be a powerful competitive tool.
To Lamb, this ability is crucial to making targeted advertising pay. Targeting broad geographic or demographic groups simply does not add enough value to make the infrastructure investment worthwhile. Household-level targeting allows advertisers, for example, to distinguish households with kids from those without – something that is not provided by Nielsen data. In practice, advertisers can break down the audience into about 200 segments for targeting purposes, he says.

[icitspot id=”9942″ template=”box-story”]

It is clear that targeted advertising still has some way to go. And if interest in the US has been slow to take off, Europe has been slower still. There are a number of reasons for this – not least  that there is simply more advertising inventory to sell in the US, where cable operators already have a large ad sales business of their own. But, as Lamb points out, there is also less demand for this type of solution in Europe because advertising is already more effectively measured (and therefore targeted), thanks to more sophisticated audience measurement systems. Lamb points out that UK advertisers have always had access to minute-by-minute ratings, unlike the 15-minute rating system prevalent in the US.
Nevertheless, TV advertising the world over will be subject to common pressures in the years to come.
“At this stage of the industry’s development there is no type of targeted advertising clearly winning out. In short, it’s too early to make any concrete predictions,” says Harmonic’s Price. “However it is not a case of ‘if’ but ‘when’ and the projections for the number of IP set-top boxes and other IP media devices clearly shows the wider adoption of targeted advertising is coming soon.”

Tags: ads