
Google is simultaneously overhauling how it ranks content, how it filters spam, and how Search Console reports AI Overview performance — all in the same week. Meanwhile, publisher referral traffic has collapsed 40% year-over-year, and agentic AI is already running enterprise networks at scale while ad executives wait for commerce to catch up.
Google’s September Spam Update Triggers Two Weeks of Volatility
Google’s September 2026 spam update is now rolling out globally, with full propagation expected to take up to two weeks — meaning rankings and traffic can swing unpredictably until it completes. Sites built on templated, scaled, or low-quality content are the most exposed, with index removals the likely outcome once the update settles.
Audit your thinnest content now and either substantially improve it or remove it before the update fully propagates.
R4T-Diffusion Changes Which Pages Feed AI Answers
Google announced R4T-Diffusion, a new production-ready query fan-out framework that determines how a single search is split into multiple sub-queries powering AI Overviews and AI Mode. Because this framework decides which pages get surfaced inside synthesized answers, a change here ripples through every AI-generated result at scale.
Start mapping which of your pages answer multi-part queries comprehensively, since those are the targets R4T-Diffusion is designed to surface.
Search Console Block Flattening Is Inflating AI Overview “Rankings”
Search Console’s block flattening method groups AI Overview data in a way that can make weak visibility look like a strong ranking position — and new reporting confirms teams relying on position metrics alone are measuring a reporting artifact, not real performance. If your AI Overview numbers look better than your traffic suggests, this is almost certainly why.
Shift your AI Overview reporting to outcome metrics — clicks, impressions, and CTR — and treat position data from Search Console as unreliable for this placement.
Publisher Search Traffic Down 40% Year-Over-Year and Still Falling
A Chartbeat-based report shows Google Search and Discover referral traffic to publishers fell 40% year-over-year — and this is an acceleration, not a plateau. Direct traffic, internal navigation, and dark social are growing to partially offset the loss, but search referral is becoming a structurally smaller acquisition channel regardless of how well a site ranks.
Build or accelerate at least one non-search acquisition channel — email, community, or direct — before search referral shrinks further.
Search Console Now Shows Traffic from Image-Based Searches
Google Search Console added a multimodal filter that reveals which pages receive traffic from image-based searches — a source that was completely invisible in standard reporting until now. There’s no query-level data yet, but this gives content teams their first read on which pages are already earning visual search demand.
Check the new multimodal filter in Search Console immediately to find any existing visual search traffic you can build on with optimized images and structured data.
Instacart Slashes Markups to Win the Grocery Delivery Race
Instacart’s CEO confirmed the company is deliberately cutting online item markups, adding loyalty perks, and introducing cheaper delivery tiers to drive grocery platform growth. It’s a lower-margin, higher-volume strategy from the category leader — and it raises the price-parity standard that every competitor will now be measured against.
If you’re in the grocery or delivery space, model the impact of a 10–20% markup reduction on your unit economics before Instacart forces the conversation.
Ad Executives Expect Agentic Buying at Scale Within a Year
IAB Europe research finds 58% of ad executives expect AI agents to autonomously purchase media at meaningful scale within 12 months — yet there is still no named live merchant handling agentic commerce at that scale today. The gap between executive expectation and actual infrastructure is the real risk, and organizations treating this as solved are likely to be caught unprepared.
Use the next 12 months to define your agentic commerce policy — what the AI can buy, at what spend threshold, and with what human override — before the infrastructure catches up to the expectation.
AI Agents Already Run Production Networks at Half of Enterprises
Cisco and Omdia research finds AI agents are actively operating production networks at 51% of surveyed organizations, with trust — not capability — cited as the main constraint on extending further autonomy. This is past the pilot stage for a majority of respondents, putting a concrete governance decision in front of every IT and operations team that hasn’t yet drawn that line.
Define your delegation threshold now: document exactly which network decisions your agents can make autonomously and which require human sign-off before an incident forces the conversation.
More from Rafal Reyzer
For deeper dives on AI and marketing strategy, visit my YouTube channel →
Hey there, welcome to my blog! I'm a full-time entrepreneur building two companies, a digital marketer, and a content creator with 10+ years of experience. I started RafalReyzer.com to provide you with great tools and strategies you can use to become a proficient digital marketer and achieve freedom through online creativity. My site is a one-stop shop for digital marketers, and content enthusiasts who want to be independent, earn more money, and create beautiful things. Explore my journey here, and don't forget to get in touch if you need help with digital marketing.