Why Your Traffic Dropped (And What It Actually Means)

If your blog's traffic chart has been sloping downward this year, the first thing worth knowing is that it isn't a story about your writing getting worse. It's a story about the biggest structural change to how people find information since the mobile internet arrived — and it happened faster, and more unevenly across niches, than most bloggers realized while it was happening to them.

This post isn't going to tell you blogging is over. It also isn't going to pretend nothing changed. Both of those are easier stories to tell than the real one, which is more useful, more specific, and — depending on what you actually write about — considerably more urgent than most general "AI changed search" articles let on.

How fast this actually happened

It helps to see the timeline laid out plainly, because the speed is easy to underestimate in hindsight:

• November 2022: ChatGPT launches and reaches 1 million users in five days — at the time, the fastest-growing consumer app in history.
• May 2024: Google begins rolling out AI Overviews broadly inside standard search results.
• Late 2024: Google's global search market share drops below 90% for the first time since 2015, as ChatGPT Search launches and directly challenges Google's core product.
• Through 2025: ChatGPT's weekly active users grow from roughly 400 million in February to around 900 million by early 2026 — more than doubling in about a year.
• By 2026: Google reports AI Overviews reaching more than 2 billion monthly users across 200-plus countries, while still holding roughly 90% of global search market share overall — meaning the disruption is happening largely inside Google's own product, not just from outside challengers (Mean CEO's 2026 research on AI search market share).

In under four years, the mechanism that used to reliably connect "someone searches a question" to "someone clicks a blog post" changed for a meaningful share of all searches. That's the part worth sitting with before getting into what it means for your specific blog.

The numbers, without the spin

A few data points capture the scale of what changed at the aggregate level:

• An Ahrefs analysis of roughly 14 billion pages found that 96.55% of blog pages get zero organic traffic from Google — a figure that predates the most recent expansion of AI Overviews, meaning the current reality is likely no better.
• Ahrefs' revised December 2025 research found AI Overviews now correlate with a 58% drop in click-through rate for the page ranking in the top organic position — up from an earlier estimate of 34.5%, meaning the effect has gotten measurably worse over time, not stabilized.
• Seer Interactive's independent analysis of informational queries found organic click-through rate collapsing from 1.76% to 0.61% — a 61% decline — with paid click-through rate on the same queries falling 68%. Notably, even queries that didn't trigger an AI Overview saw organic CTR drop 41%, suggesting user search behavior itself has shifted broadly, not just on pages where an AI Overview happens to appear (UpGrowth's 2026 vertical-by-vertical breakdown of AI Overview traffic loss).

None of this means any specific post did something wrong. It means the mechanism connecting "ranking well" to "getting a visit" weakened, industry-wide, and it weakened further and faster than most projections from even a year or two earlier.

Why the damage is so uneven across niches — and where you probably sit

This is the part general "AI changed search" articles tend to flatten, and it matters enormously for figuring out how urgent your own situation is. Industry-level tracking from BrightEdge, covering February 2025 to February 2026, found AI Overview trigger rates rising at very different speeds depending on topic area: healthcare content went from 72% to 88% AI Overview coverage, while education-related queries saw the steepest climb in the entire dataset, from 18% up to 83% (UpGrowth's 2026 vertical breakdown).

A separate ranking of consumer content categories by structural exposure to AI search found personal finance and calculator-style content facing a median 33% traffic loss over twelve months among top brands, health and symptom information down a median 43%, and recipe and cooking content ranging anywhere from 30% to 80% traffic loss depending on the specific creator (Everything PR's 2026 analysis of consumer categories most exposed to AI search).

The tech publishing sector offers some of the starkest, most concretely documented examples available. An analysis tracking ten major English-language tech publications from their traffic peak through January 2026 found a combined 58% decline in organic traffic — but the range across individual sites was enormous. Digital Trends fell 97% from its peak, ZDNet dropped 90%, HowToGeek fell 85%, and The Verge lost roughly 85% as well, while other sites like TechRadar, Wired, and CNET saw comparatively "milder" declines in the 47–74% range (Growtika's 2026 tech media traffic analysis). Notably, this pattern isn't confined to tech: NerdWallet's traffic reportedly fell 73%, and Healthline's fell roughly 50% over the same window, suggesting the exposure extends well beyond any single industry.

The common thread across the hardest-hit sites is instructive: HowToGeek's content was overwhelmingly step-by-step "how to" guides — exactly the kind of definitional, instructional query an AI Overview can now answer completely inside the search results page, with no reason for the reader to click through at all.

What actually survived — and why

Here's the detail that matters most for deciding what to do next. Not every blog in the hardest-hit categories collapsed. One analysis of the broader "blog as a business" model found that within personal finance specifically — a category with a median traffic decline near 99% for generic, description-based sites — certain sites kept growing. The distinguishing factor wasn't luck; it was whether the content described information an AI could just as easily summarize, or documented something that required the writer to actually do physical, first-hand work. Large food blogs built around original recipe development and testing, and craft-focused sites built around hands-on, demonstrated technique, were specifically cited as having weathered the shift far better than generic informational content in the same broad categories (analysis of the blog-as-a-business model through the AI search transition).

The general principle: if an AI system can compress your entire post into three bullet points without losing anything a reader actually needed, Google or ChatGPT will simply do that, and your site gets skipped. If what you've published couldn't be compressed that way — because it requires your specific testing, your specific hands, your specific documented process — the compression trick doesn't work, and the traffic mechanism holds up much better.

What this doesn't mean

A few clarifications worth making explicitly, because the doom-oriented version of this story tends to overstate things in ways that lead to bad decisions:

• It doesn't mean all blogging is dying. It means informational, easily-summarized content in heavily AI-Overview-saturated categories is under the most pressure. Original, experience-based, hard-to-compress content in the same categories has, in documented cases, kept growing.
• It doesn't mean your niche is necessarily among the hardest-hit. The spread between a 97% traffic collapse and a comparatively mild 20–30% decline is enormous, and it correlates closely with how "definitional" or "how-to" your typical content is versus how experience-driven it is.
• It doesn't mean the traffic that remains is worthless. As covered in this series' piece on the metrics worth tracking now, the visitors and citations that do come through post-shift tend to convert at meaningfully higher rates than pre-shift traffic did.

Frequently asked questions

Is this traffic decline permanent, or will it recover?
For content types genuinely answerable within an AI-generated summary — definitional "what is X" posts, straightforward how-to guides, simple comparison queries — most industry analysis treats the loss as structural rather than cyclical: the query stopped needing a click at all, and there's no indication that will reverse. For content built around firsthand experience, testing, or original data, the picture is much less settled and has shown far more resilience in the data above.

How do I find out how badly my specific blog has been affected?
Compare your traffic to individual top posts over the past 18–24 months, specifically isolating posts by content type: definitional/how-to posts versus experience-based or tested content. If you see a much sharper decline in the former category than the latter, that's a strong signal you're experiencing exactly the pattern documented across these industry studies, not something unique to your site.

Should I stop writing informational or how-to content entirely?
Not necessarily — some informational content still drives traffic, particularly where the query is specific enough that AI Overviews don't trigger reliably, or where your version includes something genuinely extractable and citable, a topic covered in this series' post on writing content AI search tools choose to quote. But it's worth being realistic: if your blog's traffic is concentrated in the exact query types shown above to be hit hardest, treating that as a wake-up call rather than a temporary dip is the more useful posture.

Where this series goes next

Understanding what changed and why it hit unevenly is the necessary first step — but it's a diagnosis, not a strategy. The rest of this series covers what to actually do with that diagnosis: which metrics deserve your attention now that pageviews alone tell an incomplete story, how to structure content so AI search tools choose to cite it rather than bypass it, and how to build the kind of direct, owned audience relationship that doesn't depend on any single platform's next algorithm update.

The traffic drop documented above is real, uneven, and in the worst-hit categories, severe. It's also, based on the surviving examples in the data, a solvable problem for bloggers willing to shift what they're actually publishing — not a verdict on blogging as a medium.