Necessity is the mother of invention. As we approach the first anniversary of the GDPR legislation in Europe, it is time to assess what impact it has had on how ads are targeted and what new technologies have emerged for reaching the right audience in a post-GDPR world.
The fears over GDPR were possibly over-hyped. The industry did not fall off a cliff — although I know of companies that have withdrawn from Europe entirely because of it — but ad spend has continued to rise (up 15% in 2018 to over £13bn in the UK alone).
What we’ve seen is a new focus on ways of targeting that does not use personal data in any way — and a renewed focus on targeting using the context of the placement, rather than the characteristics of the audience.
On the publisher side, the New York Times, Washington Post, ESPN, News US and others are doing lots of work on the impact that emotion can have on the performance of an ad. The NYT’s Project Feels takes the concept of contextual targeting (at its simplest level, put an ad for a holiday on a page that contains words like ‘Beach’ or pictures of a beach), and takes it a step further by trying to see what sort of emotion an article is likely to trigger (hope, fear, amusement). It then works out what kind of ads are most likely to work in those contexts. The NYT has used AI to develop 18 emotions or mood states that it can target ads against, and reports that it works. For example, an ad on women’s empowerment performed well against ‘self-confident’ content.
Iluma, a UK start-up, takes a different tack. Its technology places ads on websites, and once an ad works, based on metrics from an ad server, it uses AI to look for similar pages elsewhere. These criteria can include the subject matter of the article, but also the length, what style it is written in, and what the pictures show. This deliberately avoids using any data from the audience and has been shown to deliver results for categories like technology, finance, and luxury.
We are also seeing an emergence of technology that can create video ads on the fly. Personalised videos are based on contextual factors like the weather, time of day, and city-level location (not personal enough to bother the legislation). One company that does this is Spirable. They create ads that are almost like a choose-your-own-adventure story; different elements of film and text that can be put in based on the audience’s situation at the time they are seeing the ad. As with personally targeted ads, you need to be conscious of not making the ad seem too prescient, for fear of freaking out potential customers, but personalised enough to have more resonance.
We also see this approach move from online video to TV. Channel 4 now has its own version, called Dynamic TV, which takes contextual elements like date, time of day, and weather, to be able to vary the creative that is shown on catch-up on devices and smart TVs. However, this does use elements of personal data, with the user’s explicit consent. Channel 4 is lucky as there is a definite reason to ask for permission, and its viewers can see a tangible trade-off — they get ad-supported content for free, while Channel 4 gets to be able to target.
In the same way, GDPR has not really affected the biggest advertising ecosystems like Google, Facebook and Spotify. They have been able to ask for consent, and people have been willing to give it for the exclusive content and services that they are getting.
Finally, while GDPR only affects Europe so far, there is a feeling that similar legislation may spread to other parts of the world as concerns over privacy rise. It’s interesting to see the New York Times — instead of European publishers — as a pioneer in contextual targeting. Apple no longer has an ad business, so it is being vocal about how it does not track its users. Google, in its most recent developer conference I/O, announced a series of measures that will increase privacy and make it much easier to block cookies in its Chrome browser.
As we get used to a world that makes it harder to use audience-based targeting, we need to have a greater focus on using context to make messages more engaging. Tied to this, however, is the need to ensure that we are monitoring performance and can show that ads are still working.