Audience measurement by pay TV operators is now simple to implement and offers operators the possibility of increased ARPU and reduced churn. Accurate viewing data is a valuable resource that the advertising industry is willing to pay for. The picture above features Zenterio’s Audience product.
According to Magna Global, the worldwide advertising market was $512 billion in 2015. Around 28% of this total was on free-to-air TV and some 11% on pay TV – giving a total for TV of around $200 billion. Considering the large sums involved, advertisers are naturally keen to understand who is watching the programs that include their adverts.
TV audience measurement today
In most countries there is at least one company that carries out TV audience research and provides the data to the advertising industry. In many cases this company is employed by an industry association (sometimes called a joint industry committee or JIC) created by the major advertisers.
Typically audience measurement is carried out using an in-home device that detects the program that is being watched, either from its audio signature, or more recently via an embedded watermark. Data is sent from the measuring box to a central location for analysis. A back channel is provided to the household for this purpose, if not already present.
Because of the complexity of these devices, and their low volumes, the boxes are expensive. This, and the difficulty of signing-up participants, means that only a small number of households are sampled. Some small countries, for example, may be represented by only a few hundred homes. Even the USA, with 120 million TV households only measures a few tens of thousands of households. In India, with a population of over 1.2 billion, the number of measuring devices is even lower than this.
This assumes that the sample population is representative or can be effectively adjusted to be representative. In many countries the distribution of measuring devices may not be optimal. If the sampled households are all in cities for example, the differing viewing habits of the rural population, and the opportunity to sell to them, may end up being lost.
What this means is that TV viewing figures generated by these methods are only good for understanding which are the most popular programs and for measuring relative changes in viewing habits. They lack the fine detail required to track less popular programs, and are unable to pick out smaller distinct groups from the general population. For those wishing to target advertising at particular demographics through the channels and programs that they watch this presents a problem. Because of the lack of granularity, channels and programs with small, but highly targetable, audiences may actually appear to have no viewers at all.
Enlarging the TV audience sample size
According to Digital TV Research, the number of pay TV subscribers worldwide was 928 million at the end of 2015 i.e. around half of the world’s households. In many countries pay TV penetration is at over 90%, with at least 60% being typical in developed countries. So if audience measurement is carried out via the pay TV set-top box (STB) rather than a dedicated external device, sample sizes can be in the hundreds of thousands or even millions depending on the operator.
For this to work, an STB with a return path is required for data reporting. For IPTV and OTT services this is of course not a problem. Some of the other broadcast technologies such as satellite and terrestrial have had difficulty getting their viewers to connect their boxes in the past, but with the advent of VOD and catch-up services and the ubiquity of Wi-Fi this problem is fading away, at least in major advertising markets. Even if only a proportion of boxes are connected this is still an order of magnitude larger sample size than with traditional ‘metering’ solutions.
Most pay TV operators nowadays offer versions of their service that run on mobiles and tablets. So audience measurement needs to be extended to these devices as well. Then advertisers can be provided with a full picture of how many times their adverts are being seen.
Monetizing the TV data
If set-top boxes and multiscreen devices are used to collect data on viewer behavior, then the ownership of the data moves from an external market research organization to whoever manages the application on the STB or second screen. For pay TV this is the service operator.
Data directly from connected devices is more accurate and extensive than externally measured data, and is therefore more valuable to the advertising industry. As the required data only contains detail about the popularity of channels and programs, without any identifiable customer data, pay TV operators are free to sell it to the advertising industry. With margins under constant pressure this is a useful source of additional revenue.
Selling data to the advertising industry offers a quick return for minimal investment, but equally important is the benefit that the operator can gain by using the data to improve their own business performance.
Using analytics to improve the pay TV offering
Pay TV operators can combine the behavioral data with the profile information that they have on each subscriber to divide their customers into different segments depending on their viewing choices. These segments can then be used to optimize the product offering for each segment.
Content recommendations provided through the user interface can be tailored to the different preferences of the segment. Targeted mailing lists can be created with premium tiers being offered to customers who, for example, have been watching a lot of sport but have not signed up for a sports package. Customers who have not shown an interest are spared such promotions.
Overall customer satisfaction is improved reducing churn, and with higher take-up of premium content, ARPU increases too.
Using audience data to optimize carriage fees
Pay TV operators pay for the right to distribute a broadcaster’s channels. These fees are normally based on the reach of the channel, and hence it is important for an operator to have an accurate view of how many people are actually watching, so as to avoid overpaying for the service. Audience measurement provides the data.
In addition, analysis of viewing data allows an operator to understand which programs are really valued by subscribers. These may not be the ones that are the most expensive to carry. Using such insights an operator can optimize content spend, whilst at the same time improving the attractiveness of the service.
Data collection technology
For operators wishing to quickly take advantage of the benefits of subscriber behavioral data, Zenterio offers its Audience product. Audience can be linked to any source of viewing data that an operator already has. If this is not available, Zenterio can provide a data collection client that runs on set-top boxes and other connected devices such as tablets and smartphones.
Once the data is integrated, the Audience system begins to produce standard industry reports that can be used to monetize the data with third parties. Zenterio is an independent technology provider, so sales are not limited to any particular market research organization or industry body.
Set up is quick, and useful reports visualizing the data are generated from the beginning. The standard reports can immediately be used for service optimization. However, raw data is also available so that, in combination with subscriber profiles and other databases, data mining can be carried out to dig up more subtle correlations. Viewer preferences may vary with age group, location and device. Different programs may be preferred at different times of the day or day of the week and by different groups. Patterns that identify loyal customers and those that are thinking of leaving can be found.
Audience measurement holds many such insights. With Zenterio’s Audience they are now available to TV operators of any size. If you are such an operator and want to explore this further, my contact details are below.
Product Marketing Manager
+44 208 432 6143