For decades, companies have relied on performance data to understand how customers find them.
Clicks, pageviews, referral sources — these signals have shaped everything from marketing budgets to growth strategies. If the numbers looked strong, the assumption was simple: the system was working.
But that assumption is starting to break. A growing share of the buying process is now happening in places companies can’t see — and more importantly, can’t measure.
The Invisible Part of the Buyer Journey
Consider how a typical purchase decision increasingly unfolds.
A potential buyer opens ChatGPT, asks for recommendations, compares vendors, evaluates trade-offs, and receives a structured answer explaining which option is best and why. They spend 15 or 20 minutes in that interaction, forming a clear opinion.
Then they close the tab. No click. No referral. No trace in analytics.
Days later, they visit a company’s website directly, or through a branded search. To the company, it appears as a clean, high-intent visit — the kind marketers celebrate. But the most important part of that decision, the moment when preference was formed, is completely invisible.
“The most consequential part of the buying process may now happen in places companies can’t see,” says Shane H. Tepper, cofounder of Resonate Labs.
When the Data Tells the Wrong Story
This isn’t just a gap in measurement. It’s a distortion.
Research has already shown that interactions with AI-generated results are often misclassified within existing analytics systems. According to findings cited by industry tools like Ahrefs, clicks from Google’s AI-generated summaries can appear indistinguishable from standard organic traffic, making it nearly impossible to isolate their impact.
In practice, that means AI-influenced decisions are quietly absorbed into familiar-looking channels like “direct” or “organic,” giving companies a false sense of clarity about where their pipeline is coming from.
“The session that built the buyer’s conviction doesn’t exist in your measurement system,” Tepper explains. “You think you understand what’s driving performance. You don’t.”
The Companies Winning — and Losing — Without Knowing
The consequences aren’t theoretical.
In one case, a mid-market technology company appeared to be performing well across every traditional metric. Traffic was steady. Search rankings were competitive. Internally, there was little reason to question performance.
But a deeper analysis of how the company showed up in AI-generated responses revealed a different story.
While the company appeared in roughly half of relevant AI queries, it was actually recommended only a fraction of the time. A competitor with weaker underlying technology was consistently being suggested instead — not because of better performance, but because its content was structured in a way AI systems could more easily interpret and prioritize.
Nothing in the company’s analytics reflected this gap.
On paper, they were competitive. In practice, they were losing one of the most important decision points in their category.
From Measuring Traffic to Measuring Influence
The challenge is that the tools companies rely on were built for a different kind of internet.
Traditional analytics capture what happens after a click. But in an AI-driven environment, the decision-making process is increasingly happening before that moment — or without it entirely.
That doesn’t mean measurement is impossible. But it does require a different lens.
Instead of focusing solely on traffic, some companies are beginning to track a new set of signals: whether they appear in AI-generated answers, how often they’re cited compared to competitors, and how frequently they are actually recommended rather than just mentioned.
The shift is subtle but significant: from measuring visits to measuring influence.
Because in many cases, influence is now what determines whether a visit happens at all.
The Cost of Waiting
For many organizations, the instinct is to wait.
To assume that better tools, clearer attribution models, or more complete data will eventually emerge before making major changes.
But that delay carries its own risk.
“The uncomfortable part is that waiting for better data is itself a decision,” Tepper says.
As AI systems increasingly shape how buyers evaluate options, early patterns of visibility and recommendation begin to compound. Companies that understand how they are represented — and adapt accordingly — start to establish a presence that becomes harder to displace over time.
Those that don’t aren’t standing still. They’re simply not part of the conversation.