Our mission is to identify and explain the
technologies and applications that allow television services to
be provided through Internet Protocol (IP) data networks.
Readers learn the options and the system to implement IPTV along
with new features and applications and business opportunities
that are available in the IPTV industry today.
IPTV Advertising Measurements - Making More Money From Advertising
By Patrick Christian, CTO and Founder, Packet Vision
IPTV operators can make more money from advertising services by measuring viewer reactions exactly; targeting ads better; and through the use of rules based advertising. While spending on advertising increases each year with GNP, the percentage of advertising spent on traditional broadcast television and radio is declining. At the same time the percentage spent on measurable Internet advertising is increasing. According to the Television Bureau of Advertising (www.tvb.org), $10.8 billion was spent on television advertising in the United States during the 3rd quarter of 2006 which was an increase of 3.8% from the same time period in 2005. However, Price Waterhouse Coopers and the Internet Advertising Bureau (www.iab.org) report that interactive advertising grew by 36% over the same time period, reaching $4.1 billion.
IPTV technology is a pull type technology rather than a push technology (two-way communication). This means that all the media that is provided to the customer on IPTV systems is the result of commands that are received from viewing devices and these commands can be gathered to provide a very clear picture of what the viewer likes and doesn’t like.
If a large percentage of viewers regularly change the channel when a particular ad comes on, this gives a clear message that the ad isn’t working which can be fed back to the advertiser. For advertisers, this opportunity to gain valuable marketing information looks like nirvana. But how much information can they really glean, and what can they do with it?
Firstly, there are only certain things that can physically be measured on IPTV in terms of raw information. However by collecting that raw information, and processing it, very sophisticated data can be obtained.
At the lowest level, there are two key IP protocols that control activity between the set-top box and the network. One includes IGMP, which is the protocol that establishes how users select and connect to television channels. By monitoring this protocol, it is possible to see what channels people are watching, when they switch to those channels, and how long they stay on them.
The other protocol is HTTP: the protocol ‘du jour’ used by most middleware companies. In addition to the channel information mentioned above, this gives a lot of control information. For instance, with Tandberg’s middleware, every button press on the remote control results in an HTTP message. This makes it possible to monitor individual viewer activity closely from the network: even to the extent of tracking volume changes.
On a viewer by viewer - or to be strictly accurate a set-top box by set-top box - basis it can be determined which channels were switched to and when. This information can then be combined with a database that details what genre of program is being watched such as sports or cooking or gardening. A profile can be built of the types of viewer,
household by household. Therefore, from raw button press type of information, we can start to build real data that will be highly valuable to advertisers.
The same process allows us to see the length of time anyone spends looking at a particular program or piece of content, which enables interest levels to be gauged. It is important, however, to be aware of the probability that if there has not been a channel change for a long period of time it may not be that the viewer is riveted to the program, but that they have left the room. In this instance, it is necessary to work on probabilities as opposed to absolute values. The instant the channel is changed, we can be certain that there is a human being in front of the television set. As time goes on, the likelyhood of somebody being there will eventually taper off down to a probability of zero. Determining the appropriate level of tapering is not a precise science and will involve a deal of experimentation.
Profiles built from information gleaned from IP protocol activity whilst of limited use in isolation, become much more powerful tools when they are aggregated. Advertisers can then target their messages at whole audiences that meet the criteria they require rather than discrete individuals. Equally, this gets around thorny privacy issues that may well come into play if information is used on a person by person, or household by household, basis. By aggregating the data, grouping households together not only is the privacy of individuals maintained but also a better overall measurement of the audience is made since aggregation tends to smooth out statistical inaccuracies in the data that can be encountered as a result of working with probabilities.
The bi-directionality of IPTV also enables interactive advertising, which provides advertisers with a powerful way of being absolutely sure they are reaching a particular audience whether the audience is interested in the services or products on offer. If an advertisement engages the viewer sufficiently for them to press the red button for more information, that is a very positive, unambiguous indicator of a hot prospect. And once the viewer goes into an interactive application, the advertiser can offer choices that enable the general interest to be more closely defined. With car advertising, for instance, the viewer could be offered the opportunity to find out more about the performance of the car, or its comfort, or its mileage per gallon. By monitoring what the viewer selects, we can start to build a much more accurate picture of the things that are important to viewers.
Giving the viewer a menu of multiple choices that can be refined further and further in this type of interactive application effectively creates a situation where the viewers are self-profiling themselves. Through the simple process of exploring an interactive ad application the same way in which they would click through a website, a very accurate picture can therefore be developed of the type of viewer with particular reference to the product they are interested in.
Another way of profiling audiences while preserving the privacy of individuals is to use negative information. In this way, rather than trying to discover what a viewer is interested in or what refers to them,
data is collected on what they are not interested in and what doesn’t refer to them. By aggregating this type of negative information, a picture of the positive information for the desired sector of the audience can be created which effectively maintains individuals’ privacy.
Given this ability to collect data it is worth asking the cynical question – ‘Why measure anyway?’ The key to the answer is the avoidance of waste. On analysis, a huge percentage of advertising is wasted, TV advertising especially. Any system that reduces that waste is very valuable: not only for advertisers who end up paying for it but also for viewers who do not want to receive irrelevant, time-wasting ads.
Whilst the amount of advertising that can be considered irrelevant varies considerably, a recent small experiment that involved viewing a two hour movie, also tracking the advertisements put out with it, discovered that 83 per cent of the advertisers’ money was wasted and that the advertiser was wasting 17 per cent of the viewer’s time. Most viewers would have preferred to have been watching the movie rather than watching irrelevant ads. So waste is felt on both sides of the divide!
IPTV addressable advertising offers a ray of hope. Broadly speaking the categories of addressable advertising include: local ad insertion, targeted ad insertion, interactive advertising or advertising on demand
Local ad insertion is targeting by region and the smaller the geographical regions, the more targeted the advertising is. Targeted or demographic ad insertion is designed around the interests of all the types of viewers that are watching rather than their geographical location. This could be based on a wide range of factors including income groups, age, or special interests. Interactive advertising involves the one-to-one experience where the viewer has pressed the ‘red button’ on their remote control and is directly engaged with the application. All three of these types of advertising emanate through traditional, linear
television-type experiences, i.e. the viewer is watching a TV channel. The ads that are broadcasted in the commercial breaks are oriented to either the viewer’s locality or the viewer themselves, or there is an invite to press the red button.
AoD is a service that gives end users the ability to interactively request and receive advertising messages. These advertising messages can be identified from previously stored messages (advertising bookmarks) or they can be as a result of an advertising message search. Although it is still television, the viewer initiates it from the outset rather than the broadcaster and in doing so has already expressed an interest. A one-to-one link between the viewer watching television and the video server that will play out the requested content is instigated and this connection too can be used for tailored advertising. This could use any of the forms of advertising already mentioned: local ad insertion (perhaps ads for a local grocery store or restaurant); demographically-oriented advertising (perhaps based on the household’s income or whether the household includes children) or interactive. Although it is unlikely that the viewer would want to start interacting with ads at the start of a paid-for movie, this type of application could well be used at the end of the movie. In this way, VoD-type advertising translates to the linear broadcast type advertising but with a different type of control and the viewer initiating things rather than the broadcaster.
Finally, rules-based advertising is another new phenomenon that hasn’t been possible at all in the world of traditional broadcast TV but for IPTV and its ability to personalize is an ideal environment. Here, rules or policies can be created, for instance about how many ads are delivered to different viewers as well as how frequently the ads are shown and a plethora of other factors such as what time of day the ads are shown. We are then not only targeting according to interests, geography or whatever but also using advertisement consumption as a determining component. This again translates back to a better viewer experience and less waste for the advertiser.
Patrick Christian is founder and chief technical officer of Packet Vision Limited, which specializes in addressable IPTV advertising services. He can be contacted at