Digital Marketing Updates (August 17-21, 2026): B2B Benchmarks, In-SERP Buying, and Microsoft’s Max CPC Removal

AI agents are skipping company websites and trusting independent benchmarks instead, Google is testing purchases inside search results, and Meta is quietly merging your ad variations. Meanwhile, YouTube and Wikipedia are pulling in opposite directions across different AI platforms, and Microsoft Ads has a deadline you need to know about.

Here’s everything that happened this week and what it means for your marketing strategy.

01. In B2B, the Page That Wins the Sale Might Not Be One You Wrote

A piece published this week on Forbes Tech Council documents something B2B marketers need to pay attention to: when AI agents research and shortlist vendors, they are not reading company websites. They are reading independent comparison pages and benchmark tables, and they are trusting those over anything the vendor itself publishes.

One vendor cited in the piece won a deal simply by appearing at the top of a well-known, publicly checkable benchmark. The AI agent found it, verified the number was real, and moved the vendor to the shortlist. The vendor’s own marketing had nothing to do with it.

The implication is straightforward. If your company is not listed and accurate on the benchmark tables and comparison sites your industry actually uses, an AI agent evaluating vendors in your category will pass over you for someone who is. Getting on those lists, and making sure the numbers on them are current and correct, is now part of GEO strategy for B2B.

KEY TAKEAWAY AI agents trust third-party benchmarks over brand websites when evaluating B2B vendors. Get listed and accurate on your industry’s comparison tables, because the page that wins the sale might not be one you wrote.

Source: Benchmarks Are the New GEO and AI SEO Strategy for B2B SaaS Companies — Forbes Tech Council

02. Google Is Testing Purchases Inside Search Results. Is Your System Ready?

Google is running tests that let users complete bookings and purchases directly inside AI search results, without ever visiting the brand’s website. The article published this week on Inside Radio covers how this fits into a broader shift from search as a discovery tool to search as a transaction layer.

The practical risk is specific. If Google surfaces your business in an AI result and a customer tries to book or buy from inside that result, what happens next depends entirely on whether your booking or ordering system is connected and compatible. If it is not, the recommendation goes through but the conversion does not. A competitor with a compatible system gets the sale instead.

Most brands have not audited their systems against this yet. The time to check is before it goes live at scale, not after the first month of lost transactions shows up in the data. This is the kind of infrastructure question that sits at the intersection of SEO and operations, and both teams need to be in the room.

KEY TAKEAWAY Google is testing in-SERP buying. Check now whether your booking and ordering systems are compatible, or you risk getting recommended and still losing the sale to a competitor whose system actually connects.

Source: From SEO to AIO: How AI Search Is Changing Digital Marketing — Inside Radio

03. Meta Is Grouping Near-Identical Ads Together. More Variations Is No Longer the Answer

A piece in The Current this week covers a shift in how Meta’s AI handles creative volume. When the system detects ads that are functionally similar, same concept, slightly different copy or a swapped image, it now groups them and treats them as a single variation. Running more versions of the same idea does not give the algorithm more to learn from. It just recognises the duplication and compresses it.

The old logic of flooding campaigns with variations to maximise learning signals no longer holds the way it did. The system is specifically built to filter out that kind of volume now.

What actually helps the algorithm is distinct creative, ads built on genuinely different ideas, different emotional angles, different formats. A handful of those will teach Meta’s system more than dozens of minor permutations. For teams managing paid social at scale, this changes how creative briefs should be written and how many variations are worth producing per concept.

KEY TAKEAWAY Meta now treats near-identical ads as one. Stop chasing volume and focus on a handful of genuinely distinct creative ideas instead, because that is what the algorithm actually learns from.

Source: AI Made Creative Abundant. Marketers Are Rediscovering Quality — The Current

04. YouTube Dominates Google’s AI. ChatGPT Barely Uses It

A Semrush study published this week maps which content sources different AI platforms actually pull from when generating answers. The split is significant. YouTube is the most cited brand across Google’s AI systems. ChatGPT and other platforms in that ecosystem lean heavily on Wikipedia and Reddit instead.

This matters because brands and marketers often treat AI visibility as a single goal. But the platforms that drive AI answers are drawing from different wells, and a strategy built around one will leave you largely invisible on the other.

Building video content for Google’s AI and building presence on Wikipedia and Reddit for ChatGPT are not the same task, and they should not be treated as one. Content strategies that do not account for this split are optimising for one platform while staying blind on another.

KEY TAKEAWAY YouTube wins Google’s AI. Wikipedia and Reddit win ChatGPT. Build for both separately, or you will be visible on one platform and invisible on the other.

Source: YouTube Dominates AI: The Brand With the Greatest Presence Across ChatGPT, Gemini and Google AI — Merca2.0

05. Microsoft Ads Is Removing the Max CPC Cap. The Deadline Is October 1

Microsoft Ads announced this week that from October 1, new auto-bidding campaigns will no longer support a maximum CPC bid cap. The reason given is technical: the cap can actively prevent the bidding algorithm from reaching the outcomes advertisers set as goals, because it sometimes blocks bids that would have converted.

The change only affects new campaigns created after the deadline. Existing campaigns that already use a max CPC cap will continue to work as they are.

The action required now is to audit which campaigns rely on that cap as a cost control mechanism. For those campaigns, Microsoft recommends switching to Target CPA or Target ROAS bidding, which still allow spend constraints but work with the algorithm rather than against it. Leaving it until after October 1 means new campaigns go live without a control you might have been depending on. For teams running paid search, this is worth scheduling before the deadline rather than reacting to after it.

KEY TAKEAWAY From October 1, new Microsoft Ads auto-bidding campaigns will not support a max CPC cap. Audit which campaigns rely on it now and move them to Target CPA or ROAS bidding before the deadline.

Source: Microsoft Ads Is Removing Max CPC From New Campaigns — Search Engine Journal

 

---

We publish a weekly roundup of the most important updates in SEO, paid search, content strategy, and AI-driven marketing — every week, no noise.