The Crocodile Mouth Effect: Why Impressions Rise but Clicks Collapse, and What GEO Does About It
What the Crocodile Mouth Effect Is
The crocodile mouth effect describes a divergence that is now showing up across B2B marketing dashboards: impressions keep climbing while clicks, sessions, and attributable traffic fall. The two lines separate over time, forming a shape that looks exactly like what the name suggests.
This is not a short-term fluctuation or a tracking problem. It is the visible consequence of a structural change in how search and discovery work. AI summaries, featured snippets, answer boxes, and chat interfaces surface content fragments without requiring a click. Your brand is present. The user never arrives.
Understanding why this happens, and what to do about it, is now one of the more pressing questions in B2B marketing.
Why Impressions Keep Rising
Impressions rise because content is being referenced more frequently, not because more people are visiting. AI systems pull from indexed content to construct answers. Every time your content contributes to a summary, your impression count grows. The click does not follow.
Google has confirmed that a growing share of searches now end without any click, as users get what they need directly from the results page. Gotoclient’s Google Zero research puts specific numbers on this: zero-click searches rose from 24.4% to 27.2% in the US between March 2024 and March 2025, and from 23.6% to 26.1% in the EU and UK. At keyword level, the average CTR for terms triggering an AI Overview dropped from 7.3% to 2.6% over the same period.
From a visibility standpoint, brands appear more often than ever. From a traffic standpoint, they are increasingly bypassed.
Why Clicks Are Collapsing
The reason clicks fall is straightforward: the user’s question gets answered before a visit is needed. AI systems summarise options, explain trade-offs, and surface conclusions. By the time a user considers clicking through, their intent has already narrowed considerably. They are no longer exploring. They are confirming.
Gartner has documented this compression: AI-mediated discovery is shortening the early and mid stages of the buying journey and reducing exploratory browsing. The Pew Research Center found that when an AI summary appears in search results, users click through to an external site in only 8% of visits. Without a summary, that figure is 15%. That is a near-50% reduction in click-through behaviour, driven not by disinterest but by resolution.
The practical implication is that clicks now represent late-stage intent. If you are measuring early-stage influence through click data, you are measuring the wrong thing.
Why Traditional SEO Metrics Cannot Explain the Gap
SEO has always assumed a direct relationship between impressions, clicks, and value. More visibility leads to more traffic, which leads to more pipeline. That assumption is breaking down.
Impressions now measure exposure inside AI systems, not interest. A brand can appear in hundreds of AI-generated answers and receive no measurable traffic from any of them. Clicks now represent late-stage intent, not curiosity or early exploration. The metrics that used to tell a coherent story are now describing two different things.
Forrester has noted that traditional digital metrics increasingly fail to capture real influence in AI-driven journeys. The crocodile mouth is not a sign that something is wrong with your content or your SEO. It is a sign that the measurement model has not caught up with how discovery actually works.
Why This Creates a Specific Problem for B2B Teams
B2B marketing teams are under consistent pressure to demonstrate impact through traffic and pipeline contribution. When clicks fall, the internal conversation quickly turns to whether content is working, whether the channel is worth the investment, and whether visibility is actually declining.
The problem is that this conversation is often happening while something contradictory is going on in sales. Buyers are arriving better informed. They use more precise language. They ask fewer basic questions. They have already formed a view before the first sales conversation.
That gap between what the dashboard shows and what sales reports is the crocodile mouth made visible inside your own organisation. Influence is increasing. Attribution is not capturing it. The risk is making budget decisions based on the dashboard rather than on what is actually happening in the buying process.
How GEO Reframes the Problem
Generative Engine Optimisation (GEO) starts from a different premise than traditional SEO. Instead of asking how to rank higher and generate more clicks, it asks whether your brand is being accurately represented inside the AI systems that buyers now use to form their views.
The practical questions GEO is trying to answer are:
- Is our content being referenced inside AI-generated answers?
- Are we shaping how the problem is framed, or is a competitor doing that?
- Are buyers repeating our language and framing before they even speak to sales?
Instead of optimising for traffic, GEO optimises for interpretability, consistency, and reuse. In an environment where impressions are high and clicks are low, what matters is not whether someone visited your site. It is whether your positioning was present when they were forming their view.
How GEO Works in Practice
GEO narrows the gap by making content easier for AI systems to use accurately. That means building content around the questions buyers actually ask, not around the pages a site architecture suggests. It means making conclusions explicit and compressible, so an AI system can extract and cite them without distorting the meaning. It means anchoring claims with evidence that gives AI systems confidence in the source.
Concretely, this looks like: shorter paragraphs with clear topic sentences, FAQs built around real buyer questions, claims supported with data and attribution, and consistent positioning language repeated across related content so AI systems reinforce rather than contradict your framing.
When content is structured this way, impressions translate into something more durable than a session: they translate into mental availability. The buyer has encountered your framing, absorbed it, and may repeat it before they ever click anything.
Why Chasing Clicks Makes the Problem Worse
The instinct when traffic falls is to optimise for clicks: sharper headlines, more provocative previews, higher CTR at any cost. In an AI-first environment, that approach tends to backfire.
Clickbait titles and vague promises may lift CTR marginally on traditional search results, but they reduce the quality of the signal AI systems use to evaluate whether a source is worth citing. If your content over-promises and under-delivers, AI systems learn that quickly. Reuse drops. Presence in generated answers falls.
McKinsey has documented that senior B2B buyers increasingly prioritise efficiency and clarity over exploration. They want to reach a decision faster, not browse more. Content that serves that need gets referenced. Content that tries to manufacture curiosity gets ignored.
How to Measure Success When Clicks Are No Longer the Signal
If click-through rate is no longer a reliable indicator of influence, something else has to be. The signals that matter in a GEO context are different from traditional SEO metrics, and some of them require qualitative input rather than dashboard data.
Useful indicators include:
- Whether your brand appears consistently in AI-generated answers on the topics you want to own
- Whether the framing in those answers matches your positioning
- Whether sales cycles are shortening because buyers arrive more prepared
- Whether prospects are using your language unprompted in early conversations
- Whether the need for early-stage education in sales calls is declining
Forrester has noted that influence in AI-mediated journeys often manifests downstream rather than at the point of exposure. That means the measurement has to extend into sales and post-qualification conversations, not stop at the first click.
Threat or Filter: What the Crocodile Mouth Actually Exposes
Whether the crocodile mouth is a problem depends entirely on what you were optimising for.
For brands that built their content strategy around generating traffic, the gap looks like failure. Visibility is up, pipeline contribution is unclear, and the metrics that used to justify the budget are deteriorating. That is a real threat, and it requires a genuine strategic response.
For brands that were already focused on clarity, credibility, and answering real buyer questions, the gap is more like a filter. The shift to AI-mediated discovery rewards exactly the content qualities that good B2B marketing was always supposed to produce. The brands that were doing this well before AI summaries existed are finding that their content gets referenced, cited, and used. The ones that were optimising for attention are finding that attention no longer converts.
What to Do With This
The crocodile mouth is not going to close. The structural forces behind it, AI summaries resolving queries on-page, buyers arriving pre-informed, clicks concentrating at the decision stage, are accelerating, not reversing.
The useful response is not to try to restore the old traffic model. It is to build a content and measurement approach designed for the environment that now exists. That means treating impressions as a signal of presence rather than a proxy for interest, investing in GEO alongside SEO, extending measurement into sales conversations, and accepting that some of the most valuable influence your content creates will never show up in Google Analytics.
In an AI-mediated world, the question is not how many people clicked. It is whether your framing was present when the decision was being shaped.
Work with Gotoclient on Generative Engine Optimisation
If your brand needs to be visible when B2B buyers ask AI assistants for recommendations, comparisons and trusted vendors, Gotoclient helps you strengthen entity clarity, source credibility and machine-readable relevance through Generative Engine Optimisation.
Conclusion
The crocodile mouth will not close. Impressions will continue to rise. Clicks will continue to fall. The brands that win will stop treating this as a problem to fix and start treating it as a reality to design for.
GEO does not restore the old traffic model. It replaces it with something more aligned with how decisions are now made. That means building content that AI systems can use accurately, measuring influence where it actually forms, and accepting that the most important moments in the buyer journey may never appear in your analytics.
In an AI-mediated world, influence no longer flows through clicks. It flows through understanding.
Sources
- ¹ Google – Zero-Click Search Trends and AI Answers
- ² Gartner – AI and the Compression of Buying Journeys
- ³ Forrester – Measuring Influence in AI-Driven Discovery
- ⁴ McKinsey – B2B Buyer Behaviour and Decision Efficiency
- ⁵ Pew Research Center – Google users are less likely to click on links when an AI summary appears (July 2025)
- ⁶ Gotoclient – Google Zero: Now What? (2025)
