Microsoft rebuilt its bidding models and opened its agent grounding service on Bing’s index, while Google reshaped its tag and hotel campaign setups. Most performance figures today come from the companies or vendors that published them.
Microsoft’s Two Bets: Smarter Bidding And Agent Grounding
Microsoft announced a bidding rebuild on Wednesday and made Web IQ available to all Azure customers on Tuesday. Both items ship with figures that only Microsoft has measured.
- Microsoft Ads Rebuilds Its Bidding Models, Citing Gains It Alone Measured Microsoft reports average conversion value per click up 10 percent for advertisers on Maximize Conversion Value or Target ROAS, with ROAS roughly flat. The post gives no sample sizes, so the figures stay Microsoft’s own.
- Microsoft Opens Web IQ to All Azure Customers, Built on Bing’s Index Because Web IQ draws on Bingbot’s crawl, a publisher’s robots.txt, noindex and new page-level grounding preference now shape what agents can retrieve. Microsoft’s 57 percent commerce match claim comes from its own test against an unnamed rival.
Google’s Housekeeping: Tags, Hotels And Citation Cards
Google moved on three fronts, two of them documented and one only spotted by practitioners. Each asks operators to check their own setup before the schedule does it for them.
- Every Google Tag Will Become a Tag Manager Container A November 17 release note gives the Google tag Tag Manager controls and removes the gtag config command from new snippets. Google plans to upgrade unsupported installations in January 2027 but does not list which ones qualify here.
- Google Moves Hotel Campaigns Into Search Campaigns for Travel Hotel campaigns migrate in Q4 2026 into new campaigns that mirror the closest equivalent settings. Google has not said how close that match will be.
- AI Overview Citation Cards Appear Without the Fade, Spotted in Search Brodie Clark and Sachin Patel saw three citation cards at full strength beside AI Overviews, and Barry Schwartz saw a Show All button in testing. Google has confirmed none of it and has not stated a rollout scope.
Read The Source: Model Benchmarks And A Pharma White Paper
Two items rest on claims from the parties that produced them. A launch and a white paper each arrive without independent verification.
- OpenAI Ships GPT-6.2 Astra and Sol With Lower Astra API Pricing OpenAI launched both models on November 17 and prices Astra at $8 input and $40 output per million tokens. Its comparisons are OpenAI’s own figures, and nothing yet shows the launch moved ChatGPT citations.
- Pharma brand sites rank first often but earn few AI Overview citations A performance.io white paper using Semrush data finds brand sites took 9.2% of first-place rankings but only 4.2% of AI Overview citations. The authors are a vendor and its data partner, so the evidence carries that origin.
Today’s Quick Hits
- Rank Math Logs 16 Plugin Vulnerabilities in 2026, Far More Than Yoast Two remote code execution flaws rated 8.8 sit among the disclosures, and the Search Engine Journal report does not say any were exploited in the wild.