Original article was published in The Drum.

The origins of the adtech industry can be traced back as far as 1995 with the arrival of the first ad servers (I know, I was there). Three decades on, it has grown to 12 Lumascape sheets of around 4,500 adtech businesses, along with a guestimated similar number of unfunded start-ups.

It’s busy.

No wonder brand marketers recoil in horror from the barrage of adtech ‘I hope this email finds you well’ messages. They know, despite what a lot of companies say, there’s precious little chance that the solution being offered is going to shift the dial much, if at all. 

In fact, at a recent well attended London adtech event, a panel chair casually quipped that a third of all adtech companies could disappear tomorrow and it would have no effect on ad performance. I looked around and all I saw were nods of acceptance, and even approval. 

I saw Daniel Smulevich, EVP Cloud at Jellyfish, calling this out on LinkedIn recently too. He pointed out that the network benefits from extra martech ‘efficiency’ stop adding value after they reach a certain tipping point – and we’re already way past that.

Signals in the noise

So how have we got here?

Well, the graphic equaliser of media performance has been tweaked to a singularity. Any new additions are going to yield negligible to no uplift. We need some new signals in the adtech pipes to actually have an effect on performance. (Dear reader, creative data is that signal but let me get there in a moment).

Murmured in event coffee breaks for years, we’ve all known the ad ecosystem is too complex, but only now, given the triple threat of continued platform growth, holdcos in a pickle and the looming spectre of AI, have we started to call it out. 

Ciarán O’Kane, from First Party Capital, helped frame this perfectly in a recent Madtech sketch, ‘The Rise Of The Super Signal Aggregator’. These pinchpoints in the ecosystem between brands and eyeballs deliver aggregation, activation and measurement. They’re spearheading the ingestion of meaningful data signals to improve ad performance

Coding creative into a language that media understands

And so we get to creative data as the untapped super signal of the ad ecosystem. 

But what the hell is it? 

It’s a good question. Within the context of a signal to fuel the performance of the entire adtech ecosystem, it’s the ability to assess creative at massive scale. It instantly predicts how effective an ad will be based on a variety of different factors, helping marketers to decide in real time what is placed, where, when, to whom – or whether it should even run at all. 

Predictive creative data has existed before, but because of technical and logistical challenges it was incredibly slow, pertinent to only a tiny fraction of the creative flowing through the pipes of the advertising internet and not readily available in the black box of adtech. 

That’s changed, though. Finally, creative has got the technology it deserves in the form of AI-powered assessment tools like DAIVID. 

Sounds cool. What do we do with it?

First and foremost, it’s about go/no-go decisions. The poor quality of digital advertising is well documented. That was even before the ‘Sloppening’ of Gen and reGen AI ads started to wash over us with their funky extra fingers and barely brand-safe content. 

You can stop those out of the gate immediately with a score. Our own data shows 53% of the social ads in our database over index on boredom. More widely, guestimates of below par advertising that make it into the wild are already around 50%. That was an eatbigfish study, but this number is widely used as shorthand. So for a trillion dollar global industry, that’s a saving of half a billion right off the bat. Not bad. Well done, creative data. 

But that’s about raising the ‘hygiene floor’ – making sure the baseline quality meets certain standards. 

What then, though? How can we use creative data to touch the ceiling?

Scores are cool. Everyone likes scores. They are simple to understand. But just like real life, there are easy lies and difficult truths. Peel back the layers of overall creative effectiveness scores and you uncover the key components: attention, emotions, brand recall and next-step intentions.

Now consider how you could use creative data to decide how your media dollars are spent – ads with high emotion score to the top of the funnel, high next step intents campaigns to the bottom of the funnel, high first second attention prioritised for dopamine infused social and so on.

That’s beyond hygiene and helps bring about a serious reunification of creative and media planning. Touching the ceiling. 

Let’s not forget the Gen and reGen AI elephant in the room too. If I were a marketer, my first question would be, ‘does this sh*t work?’. Programmatically, creative data can decide in real time if the outputs were effective, irrelevant of being human or AI made. 

The biggest untapped signal?

The split of creative and media back in the 80s has set a long and solo journey for each. Creative has been almost studiously ignored in adtech. I get it. It’s been seen as sticky and difficult. Magical even, as Robert Beevers, Chief Effectiveness and Analytics officer of MGOMD wrote recently. 

But that’s not the case anymore. I was recently at the excellent I-COM Summit, an annual gathering of the type of smart honchos who grapple daily with these devilish details, and not only was creative data on the agenda, but there was also a real determination on the ground to prove its value. 

Not to mention live projects we’re running with our CPG, AdTech and agency partners. It won’t come as much of a surprise to know that ‘effective’ creative is increasingly being proven to drive sales through direct linkage to sales data.