How Zero-Click Search and AI Bots Hit Make-Money-Online Sites

Does an AI Overview cut clicks, or only correlate with fewer of them?

Most of the evidence on this question is observational ^[2]. A randomized study by Saharsh Agarwal and Ananya Sen tests it directly, with 1,065 US desktop Chrome users split between a control group that saw the standard results page and a second group. When an overview appeared, outbound organic clicks fell by 39.8 percent and zero-click searches rose by 34.5 percent ^[2].

Sponsored clicks did not move, which the authors read as AI Overviews substituting for organic website visits ^[2]. The study found no measurable improvement in how users rated their search experience when the overviews appeared ^[2]. The paper was first posted to the Social Science Research Network on April 3, 2026 ^[2].

How large is the drop across a wider sample?

Seer Interactive tracks click-through rate across 3,119 search terms from 42 client organizations, from June 2024 through September 2025 ^[1]. It reports CTR by whether an overview appears and whether the brand is cited in it. For the third quarter of 2025, queries with an overview that did not cite the brand had organic CTR of 0.52 percent, down 65.2 percent year over year ^[1].

Queries where the brand was cited had 0.70 percent, down 49.4 percent ^[1]. Without an overview, organic CTR also fell to 1.45 percent, down 46.2 percent ^[1].

The decline on queries with no overview matters, because it shows that AI Overviews are not the only force cutting clicks in that sample. Seer also says it cannot prove that citation causes the higher CTR ^[1]. Treat the citation gap as a pattern to test on your own pages, not as a proven lever.

Read these figures as direction, not as one industry number. The 0.52 percent CTR and the 65.2 percent drop come from Seer's client queries in one quarter, so another quarter or another site will not match them exactly. Seer's sample is broad, but it is still one agency's client mix, so weigh it against your own search data.

Why do affiliate and coupon sites feel it first?

Affiliate economics are still heavily click-based, according to Hello Partner, an affiliate marketing agency ^[3]. Coupon publishers are more exposed, because AI systems can easily surface discount codes directly inside the answer itself ^[3].

The agency cites research showing that referral traffic from traditional search engines fell over two years by 60 percent for small publishers, 47 percent for medium-sized publishers and 22 percent for large publishers ^[3]. Chartbeat figures reported by PPC Land show the same 60 percent loss for small publishers over two years ^[2].

Hello Partner also cites Ahrefs data showing that AI Overviews cut the click-through rate for position one by 58 percent in late 2025 ^[3]. The agency adds that AI Overviews are compounding damage that began with Google's helpful content updates in 2024 ^[3].

What do AI bots take compared with what they return?

Crawling is the other side of the same exchange. Cloudflare's crawl-to-referral ratio compares how many pages a platform requests with how many visitors it sends back. For June 19 to 26, 2025, Cloudflare reports that Anthropic's Claude made nearly 71,000 HTML page requests for every HTML page referral ^[4]. Mistral, at the other end of the same table, sent ten referrals for every crawl request ^[4].

Cloudflare also warns that traffic from native apps carries no referrer header, so its ratios may overstate the gap ^[4].

The ratio counts requests and visits, not revenue. Two visits can earn very different amounts, so use the ratio to spot trends across periods rather than to price your traffic.

For a make-money-online site, the chain is direct. A crawler reads your page, an answer appears on the results page in place of a click, and the visit that would have paid for the page does not arrive. Less traffic per page means less revenue per page, and the content cost stays the same.

How should you test the revenue effect on your own site?

Start with a holdout. Pick a set of similar pages, with similar traffic and revenue per visit, before you split them. Leave half of them unchanged for a month, and compare revenue per visit with the rest. The randomized design shows why a control group matters ^[2].

Keep the test window long enough to cover your normal weekly traffic cycle. Write down the change before you start, so the result cannot be fitted to what you hoped to see. Agree on the revenue definition with whoever owns the pages, so the next test starts from the same numbers.

What can the evidence not show yet?

The randomized result covers desktop Chrome users in the United States ^[2]. Cloudflare's ratios count only traffic from web-based tools ^[4].

The affiliate figures come from one agency's account. The 60 percent figure traces to Chartbeat data reported by PPC Land ^[2], so treat both as indicative. None of these sources measures a single site's revenue, so the revenue effect has to come from your own analytics.

What should MMO sites do about it?

The sources do not show which fix works best, so treat these as practical steps, not proven results.

• Measure search exposure: find the share of revenue that comes from search, and from which query groups. Hello Partner advises advertisers to watch this dependence ^[3].
• Split queries by AI Overview: separate queries that show an overview from those that do not, and by whether your brand is cited. Seer tracks queries in exactly those groups ^[1].
• Measure revenue per thousand visits, before and after each change. A click-based model can lose money while total traffic looks steady.
• Set a bot policy from data: Cloudflare recommends that creators "audit and then enforce their preferred policies for AI crawlers" ^[4]. Decide per bot, using its crawl-to-referral ratio.
• Build visits that search cannot replace: cashback publishers keep clicks because users must visit to activate a reward ^[3]. Try a calculator, a comparison or a sign-up on a small set of pages first.
• Treat AI citations as a signal, not revenue. Cited queries lost less than uncited ones, though the cause is not proven ^[1]. Hello Partner warns that citation metrics are still immature ^[3].
• Check Bing's AI reporting. Hello Partner says Bing is one of the few platforms with Webmaster tools for AI attribution ^[3].
• Rethink coupon pages. AI answers can show discount codes directly ^[3]. Keep coupon pages only where they lead to more to offer.
• Check paid traffic margins. Hello Partner says review and comparison publishers face less organic traffic and more paid spend to replace it ^[3]. Buy paid traffic only when its cost is below your revenue per visit.
• Build a direct channel, such as an email list or a community. It reaches readers without a results page. The sources do not measure how well this works for MMO sites, so test it on a small segment.

Start with the first two bullets. They show which pages are exposed before you spend on any fix.