Every marketing dashboard built in the last fifteen years runs on the same assumption: a person sees your content, clicks through, and that click gets logged, attributed, and eventually tied to revenue. Take away the click and the entire measurement stack stops working, not because the influence disappeared, but because the tool built to see it was never designed to look anywhere else.
That’s the exact problem marketing leaders are running into right now. AI Overviews and chatbots are answering questions directly, satisfying the intent that used to require a website visit, and the influence a brand had on that decision simply never shows up in analytics. The work is happening. The proof of it isn’t.
How Much of the Click Has Actually Disappeared
68% of U.S. Google searches are now zero-click, and that figure jumps to 83% for queries where an AI Overview appears, and to 93% inside Google’s AI Mode. Pew Research found that when an AI-generated summary appears on a results page, traditional result clicks fall to just 8%, and of the people who do interact with the summary itself, only about 1% click through to any of the cited source links.
The pattern holds across platforms. Similarweb’s traffic data shows AI Mode generates referral clicks at a rate of just 1.6% to 2.5% of queries, compared to 17% to 19% for traditional Google search. AI platform visits overall grew 76% year over year through the second half of 2025, while referral traffic from those same platforms plateaued. Usage is climbing. Clicks are not. That gap isn’t a measurement bug. It’s the structural signature of a search interface built to answer the question itself rather than send the user somewhere else to find the answer.
“No click. No session. No neat little attribution trail to screenshot for the board deck.”
That’s the honest version of what most marketing teams are now walking into a leadership review with, and it’s forcing a genuinely uncomfortable question: if the click is gone, what exactly are you supposed to measure instead?
Why the Old Attribution Models Can’t Just Be Patched
The instinct is to treat this as a tooling problem, add another attribution layer, connect another data source, and the dashboard will catch up. It won’t, because the gap isn’t in the tooling. It’s in where the influence is happening in the first place.
What’s often called the dark funnel, buyer activity that never generates a trackable digital signal, now accounts for an estimated 60% to 70% of B2B buying journeys. That includes AI chatbot recommendations, podcast listens, community discussions, and content consumed without ever producing a click. Branch’s research on attribution readiness found that 26% of marketing leaders still can’t trace AI-driven discovery through to conversion at all, and 24% say their existing analytics stack simply isn’t built to handle AI attribution in any form. The result shows up directly in confidence: only 41% of marketers say they can demonstrate ROI on their AI-related investments in 2026, down from 49% the year before, even as spending in the category keeps climbing.
Multi-touch attribution, the model most enterprise teams reach for as the fix, has grown to 41% adoption at the enterprise level, nearly double its 2023 rate. But only 18% of those implementations are rated as highly accurate by the teams running them, because cross-device fragmentation, walled-garden restrictions, and privacy limits all constrain what a click-based system can see, regardless of how many touchpoints it tries to stitch together.
What Actually Replaces the Click as a Signal
The teams navigating this well aren’t trying to resurrect the click. They’re building measurement around a different set of signals that describe influence rather than traffic.
Brand search volume is one of the more reliable proxies available, since a rise in direct, branded search after a period of AI-driven visibility is a strong indicator that exposure inside an AI answer is translating into recognition, even without a click at the time. AI citation share of voice, tracking how often and how prominently a brand is named inside AI-generated answers for relevant category queries, is quickly becoming the equivalent of organic rank tracking for a search landscape that no longer produces a ranked list of blue links. Self-reported attribution, something as simple as a “how did you hear about us” field on every demo request or contact form, remains one of the highest-return measurement investments available precisely because it captures influence a pixel never could.
Marketing Mix Modeling, an older, econometric approach that never depended on individual click tracking in the first place, is seeing renewed relevance for exactly this reason. And pipeline-influenced revenue, tracking whether AI-referred visitors who do arrive convert into pipeline at a meaningfully different rate than other channels, gives a directional read on quality even where the volume of trackable traffic keeps shrinking.
Rebuilding a Measurement Framework for a Clickless Funnel
- Add AI citation tracking as a standing metric, checking regularly how your brand appears when relevant category questions are asked across ChatGPT, Perplexity, and Google’s AI Mode
- Install a simple self-reported attribution field across every conversion point, since it’s the fastest, lowest-cost way to surface dark-funnel influence your analytics can’t see directly
- Track branded search volume as a leading indicator of AI-driven exposure, watching for lift that correlates with periods of stronger AI citation or media coverage
- Stop treating organic traffic decline as a performance failure on its own, and instead pair it with citation and branded search data before drawing any conclusion about visibility
- Evaluate whether Marketing Mix Modeling belongs back in your measurement stack alongside multi-touch attribution, since it was built for exactly this kind of untrackable influence
- Set expectations with leadership now that click-based reporting will keep shrinking even as demand holds steady, so a quarterly review doesn’t turn into a credibility problem later
The Shift This Is Actually Part Of
This is the same disruption already reshaping how B2B buyers research and shortlist vendors and how AI shopping agents decide what to surface to a consumer. In every case, the deciding moment has moved off a page a brand controls and into a system it doesn’t, and the businesses treating that as one connected shift, across discovery, evaluation, and now measurement, are the ones building a durable read on what’s actually working.
The click was never really the thing marketers cared about measuring. It was always a proxy for influence, convenient because it was easy to count. That proxy is breaking down, but the influence it used to represent hasn’t gone anywhere. It’s just being formed somewhere a dashboard built for 2015 was never going to see.
