If the data looks good, the strategy is working. For the last two decades, marketing performance has been built on that simple premise.
Traffic is steady. Rankings are strong. Conversion rates hold. Dashboards signal stability.
But that sense of control is becoming increasingly misleading.
According to Shane H. Tepper, cofounder of Resonate Labs, companies are still optimizing around signals that no longer reflect how buyers actually make decisions. He has been analyzing how companies show up in AI-generated responses, and why that visibility often doesn’t translate into being recommended.
A growing body of evidence suggests that companies may be optimizing for a system that no longer determines outcomes. While a different system, one that shapes decisions earlier and more invisibly, goes largely unmeasured.
And the gap between the two is starting to distort how businesses allocate budget, evaluate performance, and compete.
The Break Between Rankings and Visibility
A recent study from Semrush found that 90% of sources cited in AI-generated answers don’t come from top-ranking search results. On its own, that statistic challenges one of the most entrenched assumptions in digital marketing: that ranking highly means being visible. But the deeper implication is more disruptive.
If AI systems are selecting and synthesizing information independently of traditional rankings, then the factors that determine visibility, and ultimately influence, are no longer the same ones companies have spent years optimizing for. In practical terms, a company can dominate search results and still be absent from the answers shaping buyer decisions.
The Invisible Layer of the Buyer Journey
At the same time, the way buyers research and evaluate options is shifting. Instead of clicking through multiple websites, many are now asking AI systems to compare vendors, explain trade-offs, and recommend options. These interactions often happen entirely within the interface. No click. No session. No referral data.
By the time a buyer visits a company’s website —if they visit at all— much of the evaluation process has already taken place. What traditional analytics capture is no longer the full journey. It’s the final step. Everything that led up to that moment —the comparisons, the framing, the shortlist formation— increasingly happens in a layer that remains invisible to most measurement systems.
When Performance Data Becomes Misleading
The result is not just a blind spot, but a distortion. On paper, companies may appear to be performing well. Traffic holds. Brand searches rise. Direct visits increase. But those signals are being interpreted through a model that assumes visibility happens where measurement exists. That assumption is no longer reliable.
If AI-driven discovery is shaping decisions before a click ever occurs, then traditional metrics are capturing outcomes without context — giving credit to channels that may not have driven the decision in the first place. This creates a dangerous feedback loop. Teams continue investing in what appears to be working, while underinvesting in the systems that are actually influencing buyer behavior.
The Cost of Incomplete Information
This misalignment is not theoretical, it has practical consequences. Marketing budgets are allocated based on incomplete data. Content strategies are optimized for visibility in environments that may no longer drive discovery. And competitive positioning is evaluated without understanding how buyers are actually encountering and comparing options.
In this environment, companies aren’t necessarily making poor decisions. They’re making reasonable decisions based on partial information.
That distinction matters, because it makes the problem harder to detect. Nothing looks obviously broken, until performance starts to drift, and there’s no clear explanation why.
A Different Kind of Competition
What’s emerging is not just a new channel, but a shift in where and how competition takes place. Companies are no longer competing solely for placement in search results. They’re competing for inclusion, interpretation, and recommendation within AI-generated answers. That introduces a different set of dynamics.
Visibility is no longer just about being indexed and ranked. It’s about being selected, framed, and surfaced in ways that align with how AI systems construct responses. And those systems don’t simply retrieve information. They prioritize it, contextualize it, and, increasingly, influence the decision itself.
The Adjustment Most Companies Haven’t Made Yet
Despite these changes, many organizations continue to rely on familiar signals to guide strategy. They measure traffic. They track rankings. They optimize conversion paths. All of which still matter — but no longer tell the complete story. What’s missing is a way to account for influence that occurs before measurement begins.
Tepper’s analysis points to a broader shift already underway: one where the most important part of the buyer journey may happen outside the systems companies rely on to track it. The challenge is not that performance has become impossible to measure. It’s that the scoreboard has changed — and many companies are still playing by the old one.
In a landscape where decisions are increasingly shaped before a click ever occurs, relying on traditional data alone is no longer enough to understand what’s working.
And the longer that gap persists, the more likely it is that companies will continue optimizing for visibility in one system, while losing influence in another they can’t yet fully see.