You have probably experienced this: solid content, proper optimization, consistent publishing schedule, but organic traffic keeps declining. It is a pattern we see across our clients and the industry as a whole. The instinct is to diagnose what went wrong: maybe the headlines are not sharp enough, maybe you missed an algorithm update, maybe competitors are outpacing you.
Most of the time, though, the content is not the problem. The way you measure it is. And this distinction matters more than most marketers realize, because the wrong diagnosis leads to the wrong fix.
A SparkToro analysis shows that 68% of Google searches in the first four months of 2026 ended with zero clicks. Not 50%, not 60%. Nearly 7 out of 10. And the trend is accelerating: in 2024, the figure was 60%, so the year-over-year jump is significant. When that many searches produce no click at all, treating organic traffic as your primary performance indicator is like judging a restaurant by how many people read the menu in the window.
What actually changed
AI Overviews. That is the short answer. When Google displays an AI-generated summary above organic results, users get their answer without clicking on any website. The numbers are clear:
When an AI Overview appears, only 8% of users click on a result, compared to 15% without one. An Ahrefs study estimates a 58% reduction in clicks on the first organic result. And a randomized field experiment confirmed a 38% drop in total organic clicks when AI Overviews are active.
This shift is not a temporary fluctuation. It is a structural recalibration of how people interact with search engines. Users no longer scroll through a list of 10 blue links. They read a summary, maybe click a citation, and move on. The behavior has fundamentally changed, and our measurement tools have not kept up.
A recent article on Search Engine Journal frames the issue directly: AI content did not stop working, your metrics did. And we agree.
What Google Search Console is not telling you
Here is a problem few people are discussing: Google Search Console does not differentiate between a click on a classic organic result and an interaction with an AI Overview. You have no way of knowing if your page was cited in an AI summary, if it influenced the response, or if your brand was mentioned there.
And yet, visibility in AI Overviews matters enormously. Pages cited in AI summaries receive 120% more clicks per impression than uncited pages. Half of frequent AI Search users click on citations, compared to just 14% of occasional users. The gap is significant: your content can be present and influential in results, without any of that showing up in standard reports.
In other words, your content may be performing better than ever. Search Console is just showing you one side of the story. We explored this shift when search traffic started its decline: the numbers drop, but that does not automatically mean your brand is losing ground.
What you should measure instead
If organic traffic is no longer a complete indicator, what do you use? We have tested several approaches with our team and clients. Here is what works:
Branded search volume and direct traffic. If you publish content that does not generate direct clicks but people remember your brand and search for it later, you will see growth in branded queries and direct visits. For several of our clients, direct traffic increased even as organic declined. That is a clear signal that content influence works, just not through the traditional path.
AI surface presence. Tools like Ahrefs and SparkToro already provide data on citations in AI Overviews. They are not perfect, but they are an important starting point. If you are not measuring your content presence in AI responses, you are missing an increasingly large part of the picture. This blind spot will only grow as AI adoption accelerates.
Post-click behavior. Time on page, scroll depth, conversions, newsletter signups. If you have fewer visitors but they convert better, your content is actually working harder than before. It is the paradox of the AI era: smaller reach, but greater impact per visitor. Quality beats quantity, especially now.
A correlation dashboard. Not a single KPI, but a visualization that maps your publishing calendar alongside branded search volume, direct traffic, and conversions. When you publish a quality piece, you see the effects across these metrics even if organic traffic stays flat. Correlations tell you more than any isolated number.
Not all content is equally affected
An important distinction: not all content categories suffer equally. Pure informational pages, buying guides, "best of" articles, and content that answers direct questions are hit the hardest. Exactly the type of content AI can synthesize most easily.
On the other hand, branded searches, local queries, high-intent transactional searches, and product pages remain relatively stable. The reason is simple: users want to do something specific, not just learn something.
The strategy we recommend? Add value that AI cannot extract: interactive tools, calculators, downloadable resources, video with original perspective. Create memorable content that drives branded searches, not just ephemeral traffic. We discussed this fundamental shift before: audiences no longer search for keywords, they search for complete answers.
And one more point: do not retire pages just because traffic dropped. First check if they are cited in AI responses, if they drive branded search, or if the visitors who do arrive convert better. A Seer Interactive study showed that brand-cited AI Overview CTR dropped 61% quarter-over-quarter, but that does not mean your brand is invisible. It means you need to look elsewhere for confirmation.
Practical takeaways
Do not kill content because one number looks bad. Organic traffic is just one fragment of a much more complex picture. Check branded search, direct traffic, AI presence, and the behavior of visitors who actually reach you. The old metrics no longer tell the whole truth. But good content still works. You just need to learn to see its effects differently.
