Not every television market operates like the U.S., but all are open to similar criticism: targeting tends to be very broad; low rating shows are under-valued by sample-based measurement systems, lead times are long, cross-platform analysis of audiences is difficult or impossible and channel proliferation makes it hard for human beings to keep up with all the opportunities available.
Today, planners and buyers are mainly buying time in programs or space in a magazine or newspaper title, rather than audiences to their messages. While these programs and titles are seen and read by real people, we don’t know exactly who they are – only their broad demographic make-up. Arguably, demographics have outlived their usefulness as predictors of brand usage and will be replaced by deeper behavioural and other data.
But is this shift to audience-based buying quite as unstoppable as the proponents of programmatic buying claim? There are certainly barriers to such a transition.
Some are logistical. For example, the way TV sales are structured and organized is not conducive to the kind of ultra-flexible, near real-time trading system envisaged by the programmatic community. In the U.S. and a handful of other markets advertisers make substantial ‘upfront’ commitments to purchase airtime with performance guarantees offered based on broad demographic audience data, tying up premium inventory for many months ahead.
And while the audience ratings used to plan and buy today are not perfect, the data sets available to use in ‘audience-based’ buying are often themselves incomplete, incompatible or inaccurate. Take addressable TV advertising. A growing proportion of households in the United States are addressable through their TV set top boxes (where detailed demographic information on individual households can be combined with other databases containing address-level detail – e.g. car ownership, financial information or travel behaviour). Commercials can be delivered to individual households that fit the desired profiles and not to those that don’t.
But the extent of addressability varies by country. More than 40% of U.S. households were addressable at the end of 2016. In the U.K. around 25% are addressable through Sky’s AdSmart system. Coverage is far lower in most other European countries. Even within a country like the U.S., where it is more advanced, multiple cable operators control the market, each with their own system and approach.
More importantly, targeting a household is not the same thing as targeting specific individuals in the household. Most households have several television sets in different rooms, with plenty of choice as to what to watch. As a result, a lot of viewing within a household will be fragmented and impossible to track.
A growing number of homes are connecting their televisions to the internet – perhaps one day all TV content will be delivered in this way. As with addressable TV, households watching online can be profiled and targeted using additional data collected about them. But at this point in time, most TV viewing remains live and linear.
Trading of commercial inventory has long been based around mutually accepted audience ‘currencies’ expressed in demographic terms (e.g. Adults 18-49 or Women 16+). It is possible that this will one day be replaced by a completely new trading system where computer algorithms match inventory to buyers at lightning speed and ‘clear’ the market continuously. But nobody should underestimate the time and disruption such change will cause.
There are other factors that will slow any transition to audience-buying. Many brands actually do want to target a general audience. They know that most people will not be in the market to buy their brand at any one time. But it is equally likely that somebody will be in the market every day. So they need to keep their brand top of mind continuously, particularly when competitive brands will be out there vying for share of mind as well.
Even for products bought less frequently, like cars or insurance, targeting people who happen to have searched online for such products will likely involve plenty of ‘wastage’ – particularly if they have already bought something and no longer have any need for the product (who hasn’t been bombarded with advertising for holidays, weeks after booking a holiday?)
Many advertisers also place a value on the editorial environments in which their commercials appear. A high-end beauty product, for example, may be better fitted to Vogue magazine than to a celebrity gossip title – even if there are more readers from the product’s target audience reading the gossip title.
There is also the issue of brand safety. When an advertiser buys space in a particular program or title, the environment is largely known. When a programmatic buy is made targeting an audience without regard for where that audience is, it remains a risk that the advertiser may be seen as somehow endorsing or funding a content provider it would rather not be associated with.
We should not forget the research and data angle. Audience measurement systems have long been built around robust, high quality samples which, while suffering plenty of limitations (inability to report on audiences to the smallest media brands, often based on recall and relatively slow to report) nevertheless provide a solid foundation for understanding the audiences to media brands.
Much of that discipline and high quality has yet to be transferred to the digital world, where problems of viewability, fraud and data accuracy have been well-aired. Many of these challenges are now being addressed and tackled. We can only hope that in the ‘unstoppable shift to audience-based buying,’ the importance of credible, high quality data is not forgotten.
[Event] Canadian Gaming Summit
Join Ipsos’ Paige Schoenfeld who will be co-presenting with the BCLC sharing research highlights from a study examining the market potential of Millennials. In fact, you’ll hear why the results drove the BCLC to re-think player acquisition and create new strategies around advertising, player segmentation and product strategy.