‘A cookie-less solution’ is on the front page of every pitch deck in the advertising industry these days, a badge of confidence that what you are buying won’t be defunct a year from now. But does this mean the new solutions are any better, or has the industry just kicked the privacy can down the road?
A Consensus Emerges
When Google made its cookie deprecation announcement there was a flurry of articles from trade press and media agencies about its impact, but one consistent thread in every thought piece was need for the industry needed to radically rethink how it does digital marketing.
18 months on and it seems a consensus has been reached and the saviours of the cookie apocalypse have been established:
· First party data will now be king
· Second party data relationships will help enrichment
· ID resolution solutions will help create a solution comparable to what cookies offer
· Contextual advertising will deliver precision in-market targeting
Connectivity Remains King in the post-cookie world
The industry has moved fast to reflect the new narrative. Big businesses are investing in CDPs so they have their first party data strategy in place. Every week brings us a new ID resolution or contextual solution to deliver precision consumer targeting. Clean rooms are all the rage. The industry looks ready for the switch off.
At the same time the cookie has gone through a rebrand, from being the backbone if the industry it is now positioned as an archaic piece of ad tech which no longer respected consumer demands for privacy. However, if we scratch beneath the surface, are the alternatives that the industry has focussed on any better from a privacy or accuracy perspective?
Same same, but different
When cookie deprecation was announced the ad industry found itself at a crossroads. It could either double down on its ‘single customer view’ narrative, where it continues to talk up the potential of connecting user journeys and datasets to drive audience precision, or it could look to really shake things up and try something different.
Many large advertisers had been investing large sums in CDPs even before the cookie announcement was made, excited by the promise of the single customer view and next best action marketing this tech enables. It’s hard to argue with the approach if you have the money to invest in both the technology and the team to unlock their potential.
But this had a knock on effect; if clients were focussing on ‘enriched and connected’ first party data solutions, the advertising industry as a whole would have to double down and find their own solutions that followed the same narrative. The problem was this still depended on cookies.
So rather than do something fundamentally new, the industry has just deleted the ‘cookie’ part of the old pitch decks and replaced it with unified ID/panorama ID/ID5. Clients and the ad industry were aligned, the pitch story stayed consistent and it protected the huge data investments made by networks.
The industry took the easy option. But nobody has stopped and asked whether this new way delivers on the privacy promise that started cookie deprecation in the first place.
Protecting privacy or protecting revenue?
How has the industry managed to achieve cookie-based platform connectivity without an actual cookie? Simply it is about data sharing, migrating to common identifiers like an email address, a phone number or a postcode to stitch one user to the next. To make it privacy focussed this data is encrypted first and passed across in clean rooms, where both your data can sit but you can’t see it. The data science equivalent of dating in the dark.
Reading that back, can we truly say that this approach is more privacy focussed than using a cookie? Encrypted or not, you are still using an individual’s personal data to match across the ad ecosystem. It just doesn’t feel more privacy focussed. And if it feels less privacy focussed, it’s unlikely to last. Recent comments from EU legislators hint that they aren’t convinced by the workarounds either, suggesting this may become the next privacy battleground.
If the approach delivered genuinely ground-breaking results you could understand the enthusiasm. Salesman have been promising marketers a connected data dream for years, but poor match rates, relentless modelling to fill the gaps and black box data ecosystems often meant results didn’t live up to expectations. And this was with cookies, so it is probably safe to assume things won’t get better with less tools at our disposal.
Kicking the can down the road
The response to cookie deprecation is a familiar ad industry reaction to a more regulated ecosystem. If we look back to advertising’s relationship with the tobacco industry in the 1960s, as legislation came in advertising didn’t turn its back on big tobacco, they just shifted their approach slightly to stick within the rules.
This exact topic was covered in the opening episode of Mad Men. Don Draper and co were perplexed by how they would sell cigarettes once they couldn’t make spurious health claims, until they figured out the solution:
Don: The Federal Trade Commission and "Reader's Digest" have done you a favor. They've let you know that any ad that brings up the concept of health and
cigarettes together, well, it just makes people think of cancer.
[…]
If you can't make health claims, neither can your competitors.
Client: Great, so we got a lot of people not saying anything that sells cigarettes.
Don: Not exactly. This is the greatest advertising opportunity since the invention of cereal. We have six identical companies with six identical products... We can say anything we want. How do you make your cigarettes?
Client: We breed insect-repellent tobacco seeds, plant 'em in the North Carolina sunshine, grow it, cut it, cure it, toast it—
Don: There you go. There you go. [He writes: "Lucky Strike. It's Toasted."]
Client: But everybody else s tobacco is toasted.
Don: No, everybody else’s tobacco is poisonous. Lucky Strike’s is toasted.
While obviously dramatized, this scene isn’t far off the truth in terms of how tobacco advertising reacted, and there are obvious parallels to how the industry has responded to the cookie apocalypse. The ad industry has just done a Don Draper pitch, coming up with an idea that is legally permissible and a great sell to clients.
While this might win you the pitch, and buy you a bit of time, history also tells us that these sleights of hand only last so long. It only took so long before the loopholes in tobacco advertising closed; it is not too hard to believe that the next round of legislation will spell the end of these cookie-alternative approaches too.
The alternative path: Contextual
Interest and contextual audiences have been floated as a potential precision targeting audience solution; the Google commitment to ‘topics’ suggests this might be the solution of the future. However, this has been met with criticism as the available audiences tend to be very broad and lack the targeting precision marketer’s demand.
But it doesn’t have to be too broad if done differently. If hundreds of combinations of interest and contextual environments are run in a single campaign you can understand which combinations work best, redirecting budgets towards the best performing.
And once you understand what works best in one platform this can be used in your targeting in the next. Instead of trying to connect datasets, with all its flaws and challenges in a post-cookie world, you can test within channels, understand what drives effectiveness, and you pass learnings rather than IDs across platforms to drive efficiency.
The Tack way
This is the approach that tack has developed to target customers. Tack’s AudienServe platform enables advertisers to target customers in every platform using contextual and interest-based targeting.
But unlike other contextual solutions, we aren’t running one audience at a time or optimising each platform in isolation. We are testing hundreds of contextual and interest combinations in each platform, using machine learning to understand what is working best. Not only do you then see which attributes work best, providing important consumer insight, but these insights are then used for targeting in the next platform you go live in.
Real life testing isn’t new; A/B tests have been around for years. But the tests have always been limited because they were laborious to set up, clear recommendations hard to observe and audience insight limited. And because of the complex data ecosystem any learnings have been limited to the platform they were run in, which meant testing was never considered as a true alternative for cross-platform activation.
Because campaign set up and testing is completely automated in AudienServe, with optimisation delivered algorithmically, it overcomes these traditional limitations of the traditional A/B test. For the first time, there is a multivariate testing solution to understand how to deliver precision by platform, audience, creative and campaign.
The ad industry has spent 18 months replacing the cookie with an alternative approach and in doing so has looked at it as a technological challenge to overcome rather than a legislative adjustment to protect privacy. This isn’t right. So instead of looking for alternatives, the industry should recognise that that you do not need to connect data across platforms to drive efficiency across platforms. Once it has got its head around this alternative way of thinking about digital targeting it can then give up its addiction to ID resolution, invest in solutions that make media work harder while genuinely taking consumer privacy seriously.