How much traffic Google AI Mode is referring to sites? @Similarweb reclassified its traffic from "Gen AI" to "Organic Search" on September 1st, and we can easily see how much Gen AI traffic drops on September 1st for any site. For many of the top sites of the web, also used as citation sources, such as YouTube, Reddit, Facebook, it has been a 60-70% drop in Gen AI traffic (which is still a fraction of organic search). If you have access to Similarweb, go and take a look 😉

📊 More insights into ChatGPT’s August citation shifts via @Similarweb ! We already knew about the Reddit and Wikipedia declines, but 1) Where did this citation share go and 2) Did the shift look the same across industries? Similarweb shared with me this additional analysis of US desktop data, comparing August vs July 2026 across all prompts, as well as Beauty, Finance, and Travel ones. A few findings that stood out: * Across all 4 segments, Social Media, Reference Materials, and News & Media were the biggest 3 declines. This includes Reddit and Wikipedia. * For the 3 industry segments, the biggest increases were often in category relevant site types. E.g. Travel prompts saw Travel sites become a lot more frequently cited. This means brands in that category have a bigger chance of being cited * Citations are no longer dominated by a few UGC sites and moving towards category specific sites Take a look at the data here 👇 Thanks so much, @AdelleKehoe for the insights 🙌

I’m excited to join @randfish and @AdelleKehoe from @Similarweb in the “Beyond the Mention: From AI Visibility to Downstream Impact” webinar going beyond the usual AI visibility tracking discussion to cover what matters next: how to connect AI Search presence with actual downstream impact. We’ll be discussing: * Why the “zero-click future” is much more nuanced than it’s often presented * How AI Search visibility can influence visits even when the click doesn’t happen immediately * Where AI-influenced traffic is showing up and why it’s often misattributed * How to start connecting AI visibility with business outcomes in a more actionable way This is one of the areas I’m seeing more companies struggle with right now and that’s the gap we’ll focus on closing. Join us on July 16 👇 similarweb.com/corp/webinars… Recording will be sent to everyone who registers.

Are affiliate sites really "dead" because of AI Overviews? Not all of them! For example: RunRepeat, Pack Hacker or CleverHiker are still growing YoY, even though AI Overviews now appear on 81–95% of their top non-branded queries based on @Similarweb traffic data. A few insights, patterns and learnings: 1. AI Overviews have a high presence in their SERPs and they're still growing but... they're far more substitutive when the user only needs a quick answer, and much less so when a confident purchase decision is needed. Looking at each site's top 100 non-branded keywords, an AI Overview is present on: RunRepeat: 95%, Pack Hacker: 81%, CleverHiker: 81%. Keywords with an AIO are half to two-thirds of their estimated organic clicks... and yet, organic search traffic YoY: CleverHiker: +70% Pack Hacker: +28% RunRepeat: +6% (at a much larger scale, nearly 1M additional organic visits) 2. They're still capturing product discovery demand from Google with a high share of non-branded search traffic: Pack Hacker: 91.9%, CleverHiker: 98.1%, RunRepeat: 91.2%. They're not growing just because people already know them. 3. Their visibility is concentrated in purchase decision queries per Similarweb's intent classification: Pack Hacker: 72.7%, RunRepeat: 71.0%, CleverHiker: 64.9%. These are queries where users want to compare, validate, check real experience, understand trade-offs ... and then buy (rather than a quick answer). 4. Their top pages are decision support ones, helping users make a confident product decisions: * CleverHiker: ~87% of organic landing page clicks go to tested "best X" guides and category pages. * Pack Hacker: Roughly a 50/50 split between buying guides and reviews. * RunRepeat: ~70% goes to individual model review pages: lab-tested, measured, scored. These pages provide in-depth, first-hand insights, a variety of images and many times videos as well. 5. The three sites show more keyword losses than wins, and fewer keywords holding positions 1–3 than a year ago: Their traffic is growing while their ranking keywords decrease. The shallow long tail is eroding (that's where AI answers impact), and traffic is concentrating into fewer, stronger decision support pages. 6. The strongest sites are also building brand recognition, repeat visits and audience loyalty: the 3 sites are also building demand beyond Google, with an important direct traffic YoY growth: CleverHiker: +76% RunRepeat: +24% Pack Hacker: +23% For example, Pack Hacker also has a community/membership program with access to exclusive content and better deals. 7. The above shows that affiliate sites that have become trusted product research destinations with first-hand testing, real reviews, useful comparisons, strong UX and a recognizable brand; are still growing, even with AIOs on top of their money queries. The TL;DR: The future of affiliate SEO is becoming the site users still want to click after seeing the AI answer because a summary isn't enough.

🚨 Fantastic research from @Similarweb , with @randfish on The Downstream Impact of AI Visibility, showing something that’s critical for marketers to understand: AI influence can happen without a click, and this is why measuring AI Search impact only through “AI referral traffic” is not enough. The study found that: * Brands recommended by ChatGPT were 2.5x more likely to receive a visit within the following 7 days vs. competing brands that were not recommended. * Search was the dominant downstream channel for AI influenced visits: 55.9% of AI-influenced visits came through Search, meaning users often remembered the brand recommended by AI and later searched for it. * Your branded SERPs are now even more important: If AI-influenced demand is coming back through Search, you need to make sure your branded, brand + product, and brand comparison results are well optimized, protected, and aligned with what users saw in AI answers. * AI-influenced users engaged more deeply once they reached the site: They viewed almost twice as many pages and spent around twice as long on site compared to non-AI-influenced visitors. This is the key point: Invisible brands don’t only miss the AI mention but risk losing the downstream visit! The tricky part is that most of these visits won’t necessarily appear in analytics as “AI traffic”, making AI’s impact easy to underreport. Our current attribution models have a blind spot: AI-influenced demand often arrives through Search and Direct, not through AI referrals. So, what should marketers do? ✅ Don’t measure AI impact by referral traffic alone ✅ Treat AI visibility as a leading indicator of demand ✅ Protect and optimize your branded SERPs ✅ Connect AI recommendations with later search, traffic, engagement, and conversions ✅ Benchmark AI visibility against competitors, not only in isolation See the research here: similarweb.com/corp/the-down…