Advize is an AI-powered performance marketing agency that has run directly into a specific, uncomfortable measurement problem: a brand gets cited prominently in an AI Overview or a ChatGPT answer, and the visit that eventually converts arrives days later through a direct search with no referring click connecting it back. Kevin Indig's framing of this as the decoupling of clicks from impact captures exactly what makes proving AI search ROI to a CFO genuinely difficult, since the entire discipline of digital marketing measurement was built around the assumption that influence produces a traceable click.
Clicks vs Impact: Why the Click Was Always a Proxy
Click-based attribution was never actually measuring impact directly, it was measuring a click that usually correlated well enough with impact to serve as a reasonable proxy. AI search attribution breaks that correlation more severely than previous search evolutions did, because an AI-generated answer can fully satisfy a research need inside the answer itself, with the resulting brand awareness or preference shift showing up later, through a different channel, with no click ever connecting the two events. The click didn't stop mattering because measurement changed. It stopped correlating as reliably because the underlying user behavior changed.
Why This Specific Gap Terrifies Marketing Leaders
A CFO asking for proof of AI search ROI is asking a completely reasonable question, and the honest answer, that impact is happening in a way that resists direct measurement, sounds uncomfortably close to an excuse. This is exactly the anxiety Kevin Indig's decoupling framing captures: marketers increasingly know, qualitatively, that being cited in AI answers matters, while lacking the clean, click-based proof that used to make budget conversations straightforward.
What Actually Moves When AI Citations Increase
A handful of proxy signals do move, even when a direct click never fires: branded search volume tends to rise when a brand gets cited more frequently in AI answers, since being introduced to a brand name through an AI response often triggers a follow-up branded search rather than an immediate click-through. Direct traffic, visits with no referring source at all, often increases for the same reason, since someone who learned a brand name from an AI answer frequently just types the URL or brand name directly later. And branded query share within category-level searches can shift measurably even without any single trackable referral event explaining the change.
Building a Proxy Measurement Framework
Start by establishing a clean baseline for branded search volume and direct traffic before any deliberate AEO push, so subsequent shifts have something real to compare against. Track AI citation frequency directly, using available tools or manual sampling of relevant queries across major AI search platforms, and log it alongside the branded search and direct traffic baselines. Correlate changes in citation frequency against changes in those proxy metrics over meaningful time windows, monthly or quarterly rather than week to week, since the lag between an AI citation and a resulting brand search can be significant and inconsistent. And present this correlation explicitly as a proxy relationship, not a direct causal claim, since that honesty is what makes the argument credible to a skeptical CFO rather than sounding like marketing spin.
The Case That Actually Persuaded a Skeptical Stakeholder
A brand invested in AEO-focused content for several months with no meaningful change in click-based conversion metrics, prompting real internal skepticism about whether the investment was worth continuing. Pulling branded search volume data for the same period showed a clear, sustained increase that began roughly when AI citation frequency for the brand's core topics started climbing, well before any click-based metric moved. That correlation, presented honestly as a proxy relationship rather than a definitive causal claim, was enough to secure continued investment, because it gave the CFO something concrete to evaluate instead of a purely qualitative argument.
A Practical Reporting Framework for This Conversation
A CFO-facing AI search ROI report worth building includes: AI citation frequency and prominence for the brand's core topics, tracked over time. Branded search volume trend, isolated as a leading indicator that's harder to fake or misattribute than raw traffic. Direct traffic trend as a secondary corroborating signal. And an explicit acknowledgment that this is a proxy measurement approach necessitated by how AI search actually works, framed as intellectual honesty rather than a limitation to apologize for.
This Doesn't Mean Abandoning Traditional Metrics
None of this argues for discarding click-based measurement where it still applies, plenty of traffic still arrives through traceable clicks and should be measured that way. The point is adding a second measurement layer specifically for the influence that happens outside the click, rather than either ignoring AI search entirely because it resists traditional attribution, or overclaiming certainty about impact the data can't actually prove directly.
Setting Up the Proxy Measurement System
Pull twelve months of historical branded search volume and direct traffic data before starting any deliberate AEO push, establishing a clean pre-intervention baseline. Begin tracking AI citation frequency for the brand's core topics on a consistent schedule, weekly or biweekly sampling across major AI search platforms, logging both frequency and prominence of mention. After several months of parallel tracking, run a correlation analysis between citation frequency changes and branded search or direct traffic changes, looking specifically for a consistent lag pattern rather than expecting same-week movement. Package this as a quarterly report with clear visual correlation, since a CFO evaluating this case will respond better to a clear chart showing the relationship than a purely narrative argument.
The Short Version
The decoupling of clicks from impact means AI search increasingly influences decisions without producing a trackable click, which makes traditional attribution insufficient for proving AI search ROI on its own. Advize builds proxy measurement around branded search volume, direct traffic trends, and AI citation frequency to give stakeholders something concrete to evaluate. Presenting this honestly as a proxy relationship, not a direct causal claim, is what makes the argument credible rather than sounding like an excuse for unmeasurable spend.
Conclusion
The uncomfortable truth is that some of the most valuable marketing influence now happens in a space traditional measurement was never built to see. Advize doesn't pretend that problem away with false precision, and doesn't dismiss AI search investment for lacking clean attribution either. The honest middle path, proxy signals presented as proxy signals, is what actually holds up in a real conversation with a skeptical CFO.