Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Advertising’s synthetic bubble is about to burst

How the industry drifted from real human signals

[Photo: Getty Images]

Profile photo of J. Brooks

Founder CEO · GlassView

In 2006, some of the most sophisticated financial institutions in the world were holding assets they barely understood.

The instruments were collateralized debt obligations, or CDOs: bundles of mortgages repackaged, rated, resold, and severed from the homes underneath them. Then came synthetic CDOs, which went further still, creating exposure to housing risk without requiring a single actual mortgage. By the time the system collapsed, finance was no longer trading on houses. It was trading on representations of houses with each layer further removed from the underlying asset.

Advertising has spent the last decade building the same structure.

A real behavioral signal enters the ecosystem: Someone visits a dealership, searches for a mortgage, or watches a product video to completion. That signal is immediately captured, modeled, extended, and repackaged. The conversion becomes a lookalike model that trains another system. Those outputs are sold, licensed, and layered again. By the time a campaign reaches an “in-market auto intender,” the original signal may be several derivations removed from any directly observed human behavior.

Everyone keeps trading the asset. Almost nobody can see what sits underneath it.

THE SCARCITY PROBLEM

The deeper parallel with finance is not abstraction. It is scarcity.

In the years before the 2008 financial crisis, global capital was searching for yield. Investors wanted returns, but there were only so many high-quality mortgages to go around. Rather than accept that constraint, the financial system engineered around it, manufacturing the appearance of new opportunity without creating any new underlying asset.

Advertising has been confronting the same problem for years. Every marketer wants efficient growth. Every board wants higher return on ad spend (ROAS). Every platform promises more precision. But there are only so many people actively in the market for any product at any given moment.

At first, performance marketing works exactly as advertised. Campaigns find the people most likely to buy, returns climb, and efficiency improves. Then the highest-intent buyers are exhausted, the easiest opportunities disappear, and growth gets harder. At that point, the industry faces a choice: Invest in creating future demand through brand building or engineer new ways to extract more performance from the same pool of buyers.

Increasingly, it has chosen extraction. Lookalike audiences, modeled conversions, probabilistic attribution, attention scores, synthetic segments. Some of these create genuine value. Collectively, they reveal an industry searching for new yield after the obvious sources have been harvested. Privacy regulation has deepened the scarcity: Cookies that tracked behavior are disappearing, behavioral tracking is harder to sustain, and the direct signals underwriting digital advertising are evaporating.

It’s a phenomenon nearly every marketer eventually encounters: ROAS rises, plateaus, and then begins to deteriorate. The system is optimizing against an ever-narrower definition of success. The metrics stay healthy while the underlying asset weakens. The industry calls it optimization. An investor would call it borrowing from the future.

THE ILLUSION OF DIVERSIFICATION

Advertising has built its own version of that loop. It includes brand lift studies, attention metrics, attribution models, and third-party validation systems. Many of these measures trained on the same behavioral proxies they are supposed to independently evaluate.

In the sub-prime housing crisis, diversification appeared to reduce risk while different assets rested on identical assumptions. Advertising’s diversification is equally illusory. Different platforms, formats, and targeting strategies often sit on the same behavioral foundations. Everyone is training against the same engagement signals, optimizing toward the same proxies, and fishing in the same pool. Campaigns look differentiated on the surface. Underneath, the industry is repeating the same bet, just at different tables.

WHAT ADVERTISING ONCE UNDERSTOOD

People still speak about the advertising industry of the 1960s with something close to reverence for good reason. Those practitioners had no programmatic infrastructure, behavioral dashboards, or optimization layers running in real time. Instead, they had an obsession with people on the other side of the message: their emotions, memory, aspirations, and fears, along with the mechanics of belonging and status that drive decisions. Imperfect tools, yes, but aimed at something real.

The industry drifted from that discipline because proxy signals were faster, cheaper, and easier to scale. Why spend months understanding human behavior when a dashboard can produce millions of data points by lunchtime? Proxy, however, is not truth, and the distance between representation and reality keeps growing.

THE RETURN TO FUNDAMENTALS

Every industry eventually reaches the point where its models detach from the assets they were built to describe. Finance learned that lesson in 2008.

Advertising is approaching that same realization. Too much spend is chasing too few measurable performance signals. As direct behavioral signals become harder to obtain, the industry keeps producing new proxies, new models, and new layers of abstraction to sustain performance expectations. The question is whether those systems remain connected to the asset they claim to represent.

The real asset in advertising has never been the click, the impression, or the attribution model. Those are observations of behavior. Before every click, purchase, search, and subscription, something happens inside a human being: attention shifts, emotion forms, memory is encoded, and motivation emerges. Every signal the industry values is downstream of those processes.

Technology now allows us to observe that response directly and at scale, pairing live brain signals with campaign exposure and real-world outcomes. That does not make programmatic systems, audience platforms, or AI-driven buying obsolete. Just as the best investors combine quantitative systems with fundamental analysis, the best marketers will keep the machinery and reconnect it to a signal that is real rather than derived.

The next era of advertising will not belong to the companies that build the most elaborate abstractions. It will belong to the companies that reconnect those abstractions to reality.

Because the human being on the other side of the message was always the asset underneath the trade.

J. Brooks is the founder and CEO of GlassView.