Why Does Google Search Feel Worse in 2026 — AI Overviews Explained

Sixty percent of Google searches now end without a single click. That number explains most of the frustration.

Since Google’s landmark I/O 2026 announcement — described internally as the platform’s biggest redesign in 25 years — search has shifted from a directory to a destination. AI Overviews now appear in roughly 50% of all queries, reaching over 2 billion users globally. The answers arrive before you’ve scrolled past the first result. And for many users, that feels like a step backward, not forward.

The question everyone from SEO professionals to casual users is asking right now: why does Google Search feel so much worse? The answer isn’t simple. It involves structural incentives, algorithmic homogenization, and a slow erosion of the diverse, human-generated web that made search useful in the first place.

Four interconnected forces are driving the decline:

• AI Overviews have produced a zero-click economy that starves publishers and degrades future content quality
• Search result pages are increasingly commercial, prioritizing aggregators and ads over independent sources
• The underlying web content has homogenized, making results feel repetitive regardless of query
• Users are actively routing around Google — and the workarounds actually work

Key Takeaways

• Zero-click searches represent 60% of all Google queries as of mid-2026, with news searches hitting 69% after AI Overviews launched, according to Similarweb data via The Next Web.
• Publishers including HubSpot, Chegg, and DMG Media have documented organic traffic losses between 49% and 89% since AI Overviews scaled — with zero revenue compensation from Google for AI-served answers.
• Google’s global search market share dropped from 92.9% in 2023 to approximately 89.6% by mid-2025 — its steepest historical decline — as alternatives like Kagi, Brave Search, and DuckDuckGo absorb displaced users.
• The root problem isn’t Google’s technology getting worse. It’s that the surrounding content ecosystem has collapsed inward, with SEO optimization, AI-generated filler, and engagement metrics pushing the entire web toward identical output.

The Structural Shift That Changed Everything

The timeline matters. AI Overviews weren’t an overnight experiment — they represent the endpoint of a gradual pivot that began with BERT in 2019, accelerated through the Helpful Content Updates of 2022–2024, and crystallized at I/O 2026.

Before 2022, Google’s primary product was a ranked list of links. Traffic was the currency. Publishers produced content, Google indexed it, users clicked through. The flywheel worked because all three parties had aligned incentives.

That alignment broke when Google started answering queries directly. The 2024 US District Court ruling confirmed Google had illegally maintained its search monopoly through exclusive distribution deals. Remedies arrived in late 2025 — but those remedies targeted distribution agreements, not the AI layer now sitting above the link results. Google’s control over both the index and the answer engine remained intact.

The economic asymmetry became stark. Google continues serving ads against AI-generated answers. Publishers receive no click, no page view, no ad impression. According to The Next Web’s analysis, global publisher traffic from Google dropped 33% between 2024 and November 2025. HubSpot lost an estimated 70–80% of organic traffic. Chegg reported a 49% decline. DMG Media documented drops as steep as 89% for specific query categories.

These aren’t edge cases. They’re the publishing industry’s new baseline.

Three Forces Making Search Results Feel Hollow

The Zero-Click Economy Is Self-Defeating

The immediate experience of AI Overviews can feel efficient. Ask about drug interactions, get a structured answer. Ask about a capital city, it’s right there. Fast. Clean.

But the speed hides a longer-term problem. The content feeding those AI-generated answers came from publishers who are now losing the traffic that justified producing it. If those publishers stop — or cut quality to survive — the training data degrades. The AI answers get worse. It’s a slow feedback loop, and it’s already started.

According to Arrisweb’s 2026 analysis, generative AI has enabled publishers to flood the web with content that’s “polished in language, shallow in explanation, and minimal in originality.” Google’s algorithms, optimized for engagement signals rather than informational depth, have consistently ranked this content competitively. The result: search pages that look authoritative but deliver less genuine insight than a five-year-old forum thread.

This approach can fail in ways that compound over time. When the AI answers get worse, users lose trust. When users lose trust, they route around Google. When they route around Google, the feedback data that trains future models gets noisier. The degradation isn’t linear — it accelerates.

Content Homogenization Is an Algorithmic Output

SEO optimization doesn’t produce diversity. It produces convergence. When every publisher follows the same best practices — target the same keywords, match the same intent signals, structure content identically — results become indistinguishable.

That’s why the Reddit-query behavior became measurable. Users routinely appending “Reddit,” “forum,” or “discussion” to searches reflects a rational response to homogenized results. They’re hunting for disagreement, personal experience, and nuance that optimized articles systematically exclude.

The structural irony: forums are trusted more, yet active forum participation is declining. A shrinking pool of contributors sustains content that a passive majority now depends on. At some point, that math stops working.

Commercialization Has Narrowed Visible Results

Beyond AI Overviews, the standard results page has shifted toward ads, shopping modules, and large aggregator sites. Independent publishers — even when producing substantively better content — face systematic disadvantage against sites with larger link profiles and ad budgets.

Industry reports show users increasingly fragmenting their discovery across YouTube, TikTok, Discord, newsletters, and niche forums. This isn’t a temporary behavior. It’s a structural rotation away from Google as a primary discovery channel. And it’s happening fast enough to show up in market share data.

Search Engine Comparison: Google Alternatives in 2026

Kagi’s paid model eliminates the ad-revenue incentive that shapes Google’s results. Brave Search’s 30-billion-page independent index means it doesn’t inherit Google’s ranking biases. DuckDuckGo, at 100M daily searches, remains the lowest-friction alternative for users who want privacy without a subscription.

None of these fully replace Google’s breadth. But for specific query types — especially research or product comparisons where ad placement distorts results — they consistently outperform. This isn’t always the answer for everyone. Power users doing broad exploratory searches may still find Google’s sheer index size hard to beat. The calculus depends on what you’re searching for.

What You Can Actually Do Right Now

For power users who need Google. Several operators still work as of June 2026. Exact phrase matching via quotes forces literal string results. site:reddit.com surfaces unsponsored human discussion. filetype:pdf retrieves primary documents rather than articles about them. intitle: narrows to pages where your keyword appears in the actual title — which filters aggressively against keyword-stuffed content farms.

To bypass AI Overviews without reinstalling a browser: append -ai to any query as a per-search workaround, or embed &udm=14 into Chrome’s custom search engine URL string for permanent AI-free results. Google has confirmed there’s no official toggle planned.

For publishers and content teams. The 33% traffic drop isn’t reversible through SEO adjustments. The structural incentive has changed. Content strategies that depend on Google organic traffic need diversification now — email lists, direct indexing in alternative engines, YouTube presence. HubSpot’s 70–80% traffic loss happened to one of the world’s best-resourced SEO operations. Assuming insulation is a mistake.

Watch for: Google’s response to antitrust pressure in Q3–Q4 2026. The DOJ remedy from late 2025 left the AI layer untouched, but ongoing proceedings could force structural changes to how overviews are monetized — or mandate publisher compensation models similar to Australia’s News Media Bargaining Code.

Where This Goes in the Next 12 Months

Google’s market share dropped from 92.9% to ~89.6% between 2023 and mid-2025 — historically fast for a platform this dominant. That trajectory doesn’t reverse easily.

The key signals to track:

• Publisher consolidation: If mid-tier publishers fold at scale, AI training data quality becomes an industry-level problem, not just a business one
• Regulatory intervention: EU and DOJ proceedings could impose publisher compensation requirements or mandate transparency in AI answer sourcing
• Alternative engine growth: Brave Search and Kagi are growing from small bases — but 50M daily searches is a real number, not a rounding error

The core tension won’t resolve soon. Google earns more from a search experience that keeps users on Google. Publishers earn from a search experience that sends users to their pages. Those interests are now structurally opposed, and no algorithm update resolves that.

The product works exactly as designed. It just wasn’t designed for you.

So the question worth sitting with: if the web’s content producers can’t sustain their operations, what does the AI train on next year?

Sources: The Next Web | Wangdoo Tech News | Arrisweb

References

1. Google’s AI search overhaul is bad news for the open web
2. Google Search Is Getting Worse- Explained - Wangdoo! Tech News Wangdoo!
3. Is Google Search Getting Worse in 2026? Many Users Think So

Photo by Markus Winkler on Unsplash