For linear TV audiences, adding Connected TV and streaming services to their home viewing experience is a huge step forward, opening up a rich range of programming which is often tailored to their tastes by recommendation engines.
For advertisers trying to reach audiences in these environments though, the opposite might be said. Most brands and agencies own huge amounts of first-party audience data which powers their open-web targeting and personalised marketing activity. But this data is largely redundant in CTV.
Although broadcasters are quickly taking steps in the right direction, most CTV buying is still based on generic topic, demographic or geographic data and is certainly not tailored to the audiences of particular brands in any way.
So brands with a sophisticated understanding of their customers, and a marketing strategy to match, are left targeting relatively blind, in a space which is highly competitive and valuable.
So how can brands use the valuable insights they’ve gained from their first-party data when targeting in CTV?
1. Think about interests, not identity
First of all, it’s time to start thinking differently about audiences and what makes them relevant for advertising. It can be argued that the tags which are hard to scale into CTV – gender, age, location, for example – are not the only markers to consider. We should also be looking at the types of content which a known customer views on the open web and the places in which they engage with advertising. We call these insights ‘contextual signals’ and they can tell us huge amounts about what interests and motivates an existing customer, even when we ignore identity. Once you start listening to contextual signals carefully, you start to think about relevance in a more lateral way.
2. Model your existing customers
Once you know what interests your existing customers, you can use this information to find new, engaged audiences who are highly likely to fit a similar interests-profile to your known user. For example, if a home decor brand finds high numbers of its customers congregating on recipe pages, then by expanding its advertising campaign into more recipe content, there’s a high chance of finding more people who are interested in home decor.
The introduction of AI and machine learning into programmatic targeting means that this content-recommendation process can now be done in real time, on the fly, for the whole duration of a campaign. No need to guess the contextual interests of your potential customers – you can react and respond to your actual customers as their interests change throughout the campaign.
3. Move your data insights across platforms
The very latest contextual-AI can expand these insights across platforms. Illuma’s AI behaviorally models the live content interests of selected brand audiences from display advertising on the open web, and uses this data to power dynamic contextual targeting on the CTV side. The result is targeted, addressable CTV targeting which is driven by live audience insights, in an environment where there is currently very little original data.
Traders working this way can now offer their clients a unique way of targeting CTV that is specific to each brand’s audience, rather than ‘off the shelf’. As web and CTV are increasingly being bought in parallel, they can also benefit from a single way of working which delivers true cross platform capability, while making full use of that all-important audience data.
Originally published in Spanish in Programmatic Mexico on 06/10/2023.
