In this week’s digital news, Google will be turning on conversion-based customer lists for Google Ads advertiser accounts beginning on Jun 17, 2026. With data processing to begin on August 18, 2026. Google have announced three significant changes to its Google Ads bidding and budgeting infrastructure. With the most consequential taking effect on August 17, 2026.
By connecting Gemini AI models to private user data, including Gmail, Google Photos, and Calendar, Google can now build a deep profile of a user’s habits and needs before a search is even initiated. Google research shows why AI-generated spam is becoming harder to catch and why content-level quality filters may no longer be enough.
Google is automatically enabling conversion-based customer lists for advertisers who already use both Enhanced Conversions and Customer Match but haven’t yet activated this feature. Activation begins 17 June 2026, with data processing starting 18 August 2026. The lists will then be available to attach to campaigns for additional audience targeting. No action is needed to opt in, but advertisers who do not want the feature must turn it off in account settings before 18 August 2026. Affected accounts should have received an email from Google.
Google has launched a beta for promotion mode, a new feature for Search and Performance Max campaigns that lets advertisers pre-schedule temporary ROAS tolerance adjustments and additional daily budget for peak trading periods. Rather than manually updating targets mid-campaign, advertisers can define a date range in advance, reducing the risk of human error during busy events like Black Friday or flash sales. It is compatible with both daily and campaign total budgets. Promotion mode is distinct from seasonality adjustments, which signal expected conversion rate changes rather than directly adjusting ROAS tolerance or spend capacity. Both tools can run in parallel.
Google’s search architecture is undergoing an evolution in how it chooses which websites it wants to serve to users. By connecting Gemini AI models to private user data, including Gmail, Google Photos, and Calendar, Google can now build a deep profile of a user’s habits and needs before a search is even initiated. Search results are increasingly tailored to an individual’s specific context, meaning that if a brand does not fit the specific “model” of that user. It is effectively filtered out, regardless of its general authority or traditional keyword rankings.
For search visibility, this transition renders traditional “ten blue links” strategy increasingly obsolete. Success now requires a consistent, entity-based presence across the entire internet so that AI agents recognise a brand as a trusted option within a user’s private ecosystem. Businesses must pivot away from generic content aimed at mass acquisition and instead focus on clarity. Using structured data and token-efficient facts to ensure they remain visible when an AI agent parses the web to construct a personalised answer.
Google researchers have published a new paper detailing a sophisticated defence system aimed at curbing the “exponential challenge” of AI-generated spam and “slop.” The study notes that traditional quality filters are failing because they are easily overwhelmed by the sheer volume of unique, low-quality variations produced by generative AI. To combat this. Google is shifting its focus from content-level analysis to infrastructure signals, using “Scalable Cluster Termination” (S-CTS) to identify groups of accounts, or “Generation Clusters”. That likely originates from the same malicious automation scripts.
This has major implications for SEO visibility. As it signals a stricter environment for sites relying on programmatic or automated content production. Google’s ability to quickly adapt using Low-Rank Adaptation (LoRA) and Automatic Prompt Optimisation (APO) means that AI-generated spam is becoming significantly harder to mask. Publishers should take this as a clear warning that automated, templated, or low-quality “slop” is no longer just a violation of guidelines but is also becoming increasingly detectable at the infrastructure level. Which could lead to widespread indexing issues for sites that fail to demonstrate human-centric value.
TikTok is now one of four retailers that together make up just under half of the US beauty market. Alongside Amazon, Ulta Beauty, and Sephora. Social commerce is the fastest-growing channel. Amazon and TikTok combined could surpass Sephora and Ulta in beauty sales this year. AI platforms are projected to drive up to 35% of e-commerce transactions within three to five years. Winning on emerging channels now requires structuring product content for AI recommendation engines rather than search keywords alone.
Read more here.
AWS is in talks to sell its Trainium AI chips to other companies for use in data centres. The company has so far resisted selling its chips as compute capacity for the current chips in its data centres are already sold out. Selling its chips to other companies may mean current customers will be left waiting. Unless AWS has figured out how to manufacture more chips, which is unlikely.
Read more here.
If you need more information on any of these stories, or you’d like support on any of your digital marketing campaigns. Get in touch with our team of experts by dropping us an email to [email protected]
