Personalised advertising is nothing new. Brands have been targeting customers in line with their behaviours, interests and intents in the digital sphere for many years. But while legislative barriers like GDPR have put in certain guardrails, the precision with which such targeting occurs continues to become ever more refined.
Consumers have also become far more attuned to how they are targeted, and, while user-led pushback has created a few headlines and column inches (perhaps most notably with the #deletefacebook movement), it has not taken hold at any meaningful level.
Indeed, from a user perspective, all GDPR has led to is a flurry of pop-ups asking for permissions — which they overwhelmingly accept without questioning or thought (95% of users accept cookies, according to Teads). In fact, acknowledging their audience’s flippancy, many sites now ask users to accept cookies with a simple “yeah, whatever” button.
But while users are happy with the quid-pro-quo, many of the ads with which we, as advertisers, (re)target our audiences are icky. Annoying. And useless. Audiences aren’t all that receptive to these. This, for me, suggests that too many advertisers are ignoring the fundamental question of advertising: Who is this ad useful for?
This is a problem because marketers see that these placements have tiny but sufficient chance of conversion, typically less than 1%. Perhaps they even see a favourable CPA/ROAS – something that is becoming easier to do as attribution modeling becomes ever more understood.
Worryingly, this forgets about the 99.8% of users who won’t see the ad as a prompt for a conversion, whose reactions will vary from ambivalence to “yuck”. The tiny slice of the pie that represents the positive impact of these placements is easily measurable, but the reaction of the huge majority of impressions is far more difficult to see. These responses don’t show up in reporting that’s designed to measure reach and sales.
The prism of usefulness
The challenge is that the alternative to accurate, data-based targeting is untargeted ads. And while users hate creepy retargeting, they hate irrelevancy just as much. As such, (re)targeted media needs to always be viewed through that prism of usefulness.
Is it useful to be showcasing to a retargeted user a suite of products to buy there and then? Could it be that a more subtle, non-sales-led placement is enough to prompt a return to the site to browse for complementary products to what they have already browsed or purchased?
Direct sales data may suggest not. This gets murkier still if there are offline stores to factor in. But it might be worthwhile looking at “softer” measures of success to understand the impact on top-of-mind awareness and intent to purchase in the future as alternatives to a linear, online sales approach.
The “usefulness” in this case would come from gentle prompts rather than pushy invitations to buy there and then. Thankfully, measuring the impact of these less forceful campaigns is easier than ever, with all the main players (Google, Facebook et al) offering survey-based solutions for measuring sentiment, awareness and intent.
So, instead of being fixated on the eventual sale with our behaviourally (re)targeted ads, it’s more important to test how different kinds of placement and messaging will affect the perception of a brand across the whole spectrum of users to which they are eventually exposed (instead of merely focusing on whether they have converted or not).
Ultimately, while personalised advertising is certainly evolving, to achieve true personalisation at scale, as an industry, we need to refocus on two things: one, how we measure how users exposed to our messaging are interacting with our content; and two, how we shift their brand perceptions to drive their propensity to purchase. Get that right, and we’ll be doing a better service for the 99%.