Ogilvy knew the whole game was three moves. Make people remember you. Put it in front of as many of them as possible. Repeat until your brand lives somewhere in the lizard brain before they need it. Byron Sharp proved it with decades of data. The industry absorbed it at conferences, nodded over corporate card drinks, and walked away like a drunk leaving his wallet on the bar.
What followed is an incentives story. A story driven by rational people who don't think they're the villain but have turned into ones anyway.
The CMO who tells the CFO that programmatic attribution is producing fictitious numbers doesn't get promoted for honesty. They get replaced by someone with a cleaner dashboard. The agency that admits its media model is 30% fraud and 25% fees doesn't win the pitch. The platform that discloses its actual viewability rates loses the spend to the one that buries them in footnotes.
Every actor in the food chain makes a locally rational decision. The aggregate result is an industry that spent fifteen years optimizing toward measurement systems that are, in large part, measuring themselves.
So everybody lies. Quietly, professionally, on a loop. Long enough that the liars stop noticing. Which is when things get genuinely upside down.
In 2017, JPMorgan Chase took a hard look at where its digital ads were landing and found them lurking in the gutters of digital media, rubbing shoulders with content no serious brand should be caught dead near. They hacked their ad footprint down from hundreds of thousands of sites to a tight, controlled list. The machine didn’t sputter. Nothing collapsed. Which meant a staggering amount of that spend had been feeding ghosts, bots, and a bloated ecosystem built on the illusion of reach. Suddenly the whole digital ad economy looked less like precision targeting and more like a dressed up high-speed shell game.
That was almost a decade ago. The industry listened, ran some working groups, and then went straight back to buying the same nonsense.
In 2023, Adalytics caught Google placing ads from over 1,100 brands on sites where video ran muted, off to the side of the screen, violating Google's own stated standards. Johnson & Johnson. Disney. The U.S. Army. Samsung. Macy's. All turning up on misinformation sites. Sephora on pages hosting pirated content. Google disputed the findings. Google also quietly refunded some customers.
By 2025, global ad fraud losses hovered around $32+ billion, with some programmatic networks running fraud rates approaching 47%. The machine that generates the garbage also generates the fees, the dashboards, the reports, and the warm sensation that something measurable is happening.
Les Binet put the rot plainly after decades in IPA data: digital attribution overestimates ROI from performance media and underestimates ROI from brand advertising. Follow the attribution data far enough and you end up doing nothing but short-term activation. The brand never gets built. Demand erodes quietly across years while the dashboard looks fine, everyone keeps their jobs, and the CMO who set the strategy is three companies deep into their next chapter before anyone notices the floor is gone.
Here's the advertising fundamentals the industry actually forgot. Advertising is what you do to people before they need you, so that when they do need you, your name arrives first, unbidden, like an old friend they didn't know they trusted. It runs on memory. The slow accumulation of associations the brain files without being asked. The whole game is mental availability, and mental availability gets built over years of consistent, distinctive work. Not a week of retargeting after someone already bought the thing.
Ehrenberg-Bass calls these distinctive assets. Colors, characters, sonic signatures, taglines. The Marlboro Man. The Intel chime. The Geico gecko. The Coca-Cola script. These aren't executions. They're the actual product of the work, the thing that makes a brain skip the consideration process entirely and reach for the familiar.
“Engagement” and the philosophical lie that advertising is a logistics and agentic problem replaced all of this. Find the right person at the right moment, launch a precision message at their skull, and call it advertising. It’s direct mail dressed up in surveillance capitalism. A carousel ad nobody remembers by breakfast racks up a thousand clicks, limps across the conversion line, and gets celebrated as success while creating precisely zero brand equity. The industry reenacts this farce every week with a straight face and billion-dollar budgets.
But wait. As if the delivery mechanism wasn't sufficient self-destruction, the industry decided to kneecap creative at the same time. US adults spend over twelve and a half hours a day with media. Screens follow them from bedroom to commute to couch. More surfaces, more eyeballs, more hours than any marketer in history ever dreamed of. The industry's response was to serve muted ads autoplaying off to the side of the screen, follow people around the web for a week after they've already bought the thing, outsource brand identity to influencers with their own brands to build, and generate AI creative at scale, a thousand variations of the same beige visual with different headlines, personalized to the individual and memorable to absolutely no one.
Almost ten years ago, Nielsen pegged creative quality at 47% of sales variance driven by advertising. The industry's response of course was to cut creative budgets and call the targeting more sophisticated.
The fix is neither complicated nor popular. Shamefully forgotten, it's book-one, page-one level. Start with brand building creative and treat it as the asset, not the decorative layer bolted onto the media plan after the real decisions are made. Not AI creative. Not mass-personalized creative. The brief isn't how do we reach the individual. It's how do we make something that works on everyone, that travels, that lodges. One clear human truth, expressed with enough distinctiveness that you know immediately who it's from, repeated until it becomes part of the furniture of the culture.
The iPod silhouettes. The McDonalds arch. The Resse's orange and yellow. Dove's Real Beauty before it got dulled down for people in expensive loafers who got nervous. None of those needed a data clean room. None were segmented to a demographic. They were made for everyone, which is precisely why everyone felt them.
Pre-test the work before it runs, with System1, Kantar, or Ipsos tools that score for long-term brand effects, not 30-day click rates. The creative that scores highest on emotional response is usually more expensive and harder to justify to a CFO who wants the targeting rationale rather than the psychological one. That's the CFO's problem to outgrow.
Bring econometric modeling back alongside attribution data. Media Mix Modeling is slower and more expensive than last-click attribution. It doesn't produce a clean story and it doesn't tell anyone what they want to hear. It also actually isolates causal effects, correctly values brand-building activity that attribution is structurally incapable of measuring, and stops the slow budget bleed toward short-termism that hollows brands out across years. The encouraging news out of MediaPost from earlier this year was how several large advertisers returned to MMM after realizing their attribution models were producing answers they could mathematically prove were fiction. That last part of the sentence should horrify people more than it does.
Then audit the supply chain with the same discipline Chase applied in 2017. Find what percentage of programmatic spend generates anything beyond a served impression. Cut the long tail without sentiment. The ANA found that just 36 cents of every programmatic dollar actually reaches a consumer. A full quarter of the open web market disappears into low-quality inventory and fraud before it touches a human being. No other industry would accept this level of loss. The advertising industrial complex celebrates it at scale.
Redirect what you recover toward reach, relevancy, repetition, and creative. That's where four decades of evidence says the returns live.
And then the someone has to walk the CFO through a model where brand investment doesn't pay back in 30 days but does pay back across three years, and hold that line while three people in the room suggest better attribution tooling as a counterproposal. Binet and Field's IPA research established 60% brand-building to 40% activation as the ratio that produces sustainable growth, across categories, across geographies, across four decades. Most brands have inverted it entirely. Most have a story about why their category is different.
The category is never different. The math doesn't care.
Until someone finally decides that the numbers on the slide are not the same thing as a brand building and business, the dashboards will keep looking great and the brands will hollow out underneath, silently decaying.
originally published: May 14, 2026