This week, Shirley looks at why ad tech is much like trying on a swimsuit in the height of summer - labels, mirrors and the big reveal...

Sun and water, the ultimate summer combination. Water and swimsuits, the necessary combination. Easily one of summer's most revealing combinations: standing in a fitting room with a 360-degree mirror view and unforgiving lighting… in swimwear.

Front, side, back, every angle becomes visible, all at once. Numerous mirrors and harsh LEDs strip away every flattering illusion and what normally stays hidden suddenly becomes impossible to ignore.

So what would ad tech in a swimsuit look like? In a fitting room that removes all flattering conditions. No campaign deck, conference stage or soft lighting from the DSP dashboard. Just bright lights and three mirrors.

The first thing it would probably notice is the label. Like swimwear brands that insist every piece is "slimming", "sculpting" or "confidence enhancing", every ad tech company seems to have settled on the same description. AI-powered. Agentic. End-to-end. Privacy-first. Omnichannel. The labels are doing an extraordinary amount of work before anyone even tries the product on.

The Mirror

Then there are the mirrors: one reflecting how the company sees itself (revolutionary), another how customers see it (potentially useful), and a third how competitors see it (a feature we announced six months ago).

The front mirror is usually reassuring. The industry looks healthier than its critics suggest and it doesn't get tired of celebrating that. The conference calendar barely has a blank space in it, and to be fair, there is a lot to cheer for. Ad tech grew from a niche within digital, back when digital itself was still a niche, into the dominant form of digital ad buying. It has solved problems and opened opportunities at scale, moving the digital advertising ecosystem into the post fax machine age. Granted, every now and then someone writes its obituary. Ad tech reads it, shrugs, and survives yet another privacy regulation, browser decision, AI revolution or economic downturn. That's not nothing.

Then you turn sideways and that's where all the extra layers become visible. A dashboard here, an identity graph there, an AI agent bolted on somewhere in between, and an optimisation layer promising to simplify everything while making everything slightly more complicated. Ad tech has become incredibly good at accessorising itself. Every weakness gets another strap instead of a redesign. Feature lists meanwhile grow longer while customer and industry issues remain stubbornly familiar. 

The Reflection

Measurement is probably the neatest example. Media Mix Modelling in particular, which has spent the past years being (re-)introduced as the moment the industry finally took its clothes off: open, neutral, and independent from platform attribution. Then the walled gardens turned up too, Google with Meridian, Meta with Robyn, both open source. Yet Google’s own version of that supposedly neutral model seems remarkably well acquainted with its own ecosystem. You start wondering whether all those layers are providing actual support or simply hiding an awkward fit.

Then comes the mirror nobody enjoys. The one behind you. Didn't we just celebrate ad tech solving problems at scale? We did. And yet, ad tech has simultaneously created just as many issues, if not more: privacy, ad tech tax, fraud, spam, scam MFAs, and the rest you’d rather not see. Underneath all of it, incentives steering the whole industry: to add another intermediary rather than remove one, to measure activity because it's easier to report than outcomes, to chase quarterly revenue ahead of long-term trust, to ship products customers have to be trained on instead of ones they understand on sight.

Again, you turn sideways: every extra layer looked reasonable in isolation, but the mirror behind shows what happens after twenty years of reasonable decisions. The cumulative effect of "one more". Nobody deliberately designed an ecosystem this complex but it still emerged from thousands of those individually sensible decisions. 

It happened one papercut at a time. Privacy regulation created a whole industry of privacy technology, fragmentation did the same for orchestration platforms, and once identity started disappearing, somebody built a graph to replace it. Every solution then generated a market to sell it, a vendor to run it, and one more layer stacked on top. None of them looked particularly excessive at the time they arrived but together they explain why simplification has become one of ad tech's fastest-growing product categories, requiring just one final layer whose job is to simplify the layers beneath it… AI.

The Reveal

Finally, there is always the question nobody wants to hear from outside the curtain.

"So how does it fit?" In the end, that’s all prospective customers really want to know. Not how many features exist, or how many billions of impressions flow through the platform. Not how many slides explain the (overwhelmingly complex) architecture. Just whether this solves their problem. It sounds like a low bar but rarely is. Which may explain why so many marketers decide to skip the fitting altogether and simply buy off the rack from Google, Amazon or other black boxes. They may not know exactly how or why, but they (think to) know black swimwear fits everyone.