The first half of September brought a lot of change across paid media and search, and several AI-driven updates that had been discussed for months suddenly became realities for marketing teams.
Google began moving a large number of Search campaigns to AI Max. Meta started removing some placement controls from Ads Manager. ChatGPT sharply reduced its use of Reddit as a cited source and also updated its advertising rules. Amazon expanded Sponsored Products into creator content, while the IAB released new guidance for measuring visibility in AI-powered search. Meanwhile, Nielsen’s acquisition of DoubleVerify raised new questions about the future of independent ad verification.
Here’s a look at the biggest changes from September 1st–15th, what they mean for digital marketers, and where teams should focus their attention next.
Google Ads Started Auto-Migrating Search Campaigns to AI Max on September 1st
Google Ads began automatically moving eligible campaign-level Broad Match and legacy Automatically Created Assets campaigns to AI Max for Search. The rollout is happening throughout the month, so advertisers don’t need to change a setting for the migration to take effect. Brand inclusions and exclusions carry over.
This was already in motion. Google stopped allowing new campaign-level Broad Match and ACA entities in the UI, Editor, and API on August 3. The September rollout is essentially the final stage of that transition.
There’s also another deadline ahead as Google is encouraging advertisers to move Dynamic Search Ads to AI Max before a separate forced migration planned for February 2027, when creating new DSA ad groups will no longer be possible.
For marketers, the immediate job is straightforward: identify campaigns scheduled for migration and establish a performance baseline from before September 1 (hopefully you’ve already done that). Track CPA, conversion rate, impression share, and top-performing search terms at the campaign and ad group level. Doing this will give you something to compare against once AI Max has been running for several weeks.
It’s also worth reviewing negative keywords, search-term exclusions, and brand exclusion lists. Those controls still work with AI Max and may become even more important as Google takes a larger role in deciding which searches and assets are used.
Microsoft’s Approach
Microsoft moved in the opposite direction just days earlier by making AI Max generally available and enabling it by default for new Search campaigns. Its system combines search-term matching, text customization, and final URL expansion.
Microsoft reports at least an 8% conversion lift based on 44 advertiser A/B tests conducted between June and August 2026. The spend-weighted lift was about 13.6%, with a lower confidence bound of 8.2% across 10 statistically significant tests. Those figures come from Microsoft’s own testing, so they’re useful context rather than independent evidence.
One notable difference is Dynamic Search Ads. While Google is phasing DSA out, Microsoft says it will continue supporting the format for now. It has also added reporting for AI-optimized matching, AI assets, landing pages, and term exclusions.
For advertisers frustrated by the lack of visibility in some automated campaign products, Microsoft is offering more reporting detail. There’s one other deadline to keep in mind: Max CPC will disappear from new non-portfolio Microsoft campaigns on October 1.
No September Core Update Was Announced, but Search Kept Moving
Google did not announce a broad core update during the first half of September (We can hear the cheers from SEOers everywhere). The two confirmed 2026 core updates were the March and May rollouts, with the May update finishing June 2.
Ranking volatility remained high, with the effects of Google’s August 2026 spam update still showing up. The update ran from August 18 to August 21. SE Ranking analyzed 100,000 U.S. keywords across 20 industries and found that 16.71% of URLs previously ranking in the top 10 fell below position 100, compared with 9.2% during its comparison period. That represented an 82% increase in top-10 volatility.
Google’s recovery advice hasn’t changed: give the system time. Ranking changes can take months to settle, and there is no separate “helpful content” system to optimize for anymore; those signals are now part of Google’s broader ranking systems.
For SEO teams, the bigger mistake is looking only at the sitewide traffic chart. A better approach is to compare the two weeks before August 18 with the two weeks after, breaking the data down by page, query, directory, and page type. Then look at which URLs gained visibility instead of yours. The cause could be content quality, a competitor’s stronger page, or even a technical issue that existed months earlier but became more important after the algorithm changed.
Two Other Search Changes Worth Watching…
Google also acknowledged a Search Console data issue affecting generative AI reporting. An impressions logging bug began August 13 and continued into early September. So, if AI Overview or AI Mode impressions appear to drop sharply in mid-August, check the reporting issue before assuming search visibility actually declined.
Google also confirmed that some AI Overviews can now expand into much longer, AI Mode-style answers without requiring users to click “Show more.” That takes up more space, especially on mobile, and changes the value of traditional organic rankings whenever an AI Overview appears above the results.
Gmail Search AI Overviews Went Global
Google rolled out AI Overviews in Gmail search globally in September, and the feature works much like Google’s search experience but applies it to a user’s own inbox. Rather than delivering matching emails, Gmail can now provide a summary of relevant threads that includes citations referencing the original messages.
Earlier this year, Gmail also introduced automatic thread summaries and personalized reply suggestions, bringing AI interaction when using email to the forefront.
For marketers, AI will often digest subject lines and preview copy long before human eyes see them. Because these automated tools distill messages down to their core intent, cryptic or hyper-dramatic hooks risk getting flattened in summary, whereas straightforward, clear writing ensures the core message survives intact.
Meta Started Removing Ad Placement Exclusions from Ad Sets
Rather than allowing ad sets to outright blacklist specific platforms or placements, Meta is steering media buyers toward value rules; a mechanism that lets campaign managers scale placement bids anywhere between +1,000% and -90. Account-level placement exclusions are still available through Advertising Settings and take priority over value rules.
Meta says Advantage+ placements deliver about 11.7% lower cost per action than manually selected placements, based on its own data. The bigger question for advertisers is whether lower average costs are worth giving up control over where ads appear.
The change follows a broader pattern of Meta reducing placement and targeting controls. Detailed targeting exclusions were reduced in 2025, while additional placement restrictions followed throughout late 2025 and 2026.
For teams managing Meta campaigns, the immediate priority is to move any important exclusion rules to the account level while that option remains available. Creative should also be reviewed with broader placement delivery in mind. If certain placements consistently underperform, build value-rule templates for those situations and document the logic so bids can be adjusted as performance changes.
ChatGPT Stopped Citing Reddit Almost Overnight
One of the more noticeable recent changes in AI search happened inside ChatGPT. Reddit’s share of ChatGPT Search citations fell by roughly 86% over four days in mid-August, dropping from about 3.83% to 0.52%. The decline continued into September, with several independent AI-visibility platforms reporting a similar pattern.
OpenAI has not publicly explained the change. One theory, based on analysis of ChatGPT’s search behavior, is that the system may now identify specific entities before searching the wider web. For example, a query about AI note-taking apps could prompt ChatGPT to look for known products first, potentially reducing the role of large community sites such as Reddit.
For teams working on answer-engine optimization, the shift is another reminder that AI-search visibility can change quickly. Strategies built heavily around Reddit citations may not produce the same results they did a few weeks ago.
Brand recognition may also matter more. If an AI system already knows a company or product well enough to include it in its initial search, that brand may have an advantage over newer competitors. It’s another reason to look beyond individual backlinks or publication placements when measuring visibility in AI-generated answers.
ChatGPT Ads Kept Expanding, and OpenAI Added a New Advertising Rule
ChatGPT Ads expanded to 31 European markets following its August rollout. OpenAI is now testing two campaign objectives: the original reach objective and a newer click objective, while self-serve access continues to roll out in stages.
OpenAI also introduced a notable restriction: ads from competing brands cannot appear inside a conversation about another brand. For a relatively new advertising platform, that is an unusual limitation, particularly as OpenAI continues expanding its ad inventory.
The policy could help maintain user trust, but it also limits the types of competitive advertising the platform can sell. More broadly, it may set expectations for how advertising works on other conversational AI platforms.
OpenAI has also reported that ads per user per hour on the ChatGPT mobile app in the U.S. increased by an average of 26% per month between April and August. That suggests ad inventory is moving beyond a small-scale experiment and becoming a more established part of the product.
For marketers, the bigger story is the intent signal. Someone asking an AI a detailed question about a product or category may be further along in the buying process than someone entering a short search query. Early advertiser reports have pointed to lower acquisition costs than comparable Google Search campaigns, although those results are still emerging and come from advertisers using the platform.
As more brands test ChatGPT Ads, competition for that inventory could increase significantly.
Amazon Sponsored Products Now Serve Inside Creator Content, and Campaigns Were Enrolled Automatically
Amazon expanded Sponsored Products into content from creators in its Amazon Influencer Program on August 10. Now, the important detail for advertisers is that existing Sponsored Products campaigns were automatically enrolled and Amazon notified advertisers that no action was required, and campaigns carried their existing bids and budgets into the new placement.
However, there are two changes marketers need to watch.
First, Top of Search and Product Pages bid adjustments do not apply to off-Amazon placements, although dynamic bidding still does. Basically, what that means is carefully optimized bid modifiers may no longer work when an ad appears in creator content.
There’s another wrinkle, too: creator content usually isn’t tied to an actual search query. Amazon will still assign an inferred search term and drop it into your search-term report, which muddies negative-keyword analysis, since some of those terms were never actually typed by the shopper.
You do have some levers. Limiting off-Amazon spend and excluding individual creators are both options, but Amazon hasn’t given advertisers a clean, account-level switch to turn the placement off entirely. By default, the setting is Increase reach, so if you haven’t touched anything, you’re opted in.
None of this is surprising given where Amazon’s ad business is headed. The company pulled in $19.8 billion in ad revenue in Q2 2026, up 26% year over year, and that’s a lot of incentive to push Sponsored Products into whatever new inventory it can find.
So, what should you do? Run a quick audit. Pull your search-term data from the two weeks after August 10 and flag anything generic or poorly matched. If creator placements don’t make sense for your brand, flip on Limit off-Amazon spend. And go back to your campaigns where placement-specific bid modifiers were carrying a lot of weight; they may need adjusting now that the landscape has shifted.
The IAB Published New Guidance for Measuring Visibility in the AI Era
The IAB’s new Measuring Visibility in the AI Era report is becoming an important reference point for marketers trying to understand how their brands appear in AI assistants. Released in late August as part of Project Eidos, the roughly 36-page framework outlines things like core metrics, measurement standards, provider disclosures, and guidelines for making results more stable and reproducible.
One of its main ideas is the “four Ps”: presence, prominence, portrayal, and persuasion. The IAB also separates measurements into directional and decision-grade data, depending on how reliable and actionable the results are.
That distinction matters because AI-visibility measurement is still far from standardized. The IAB says only about 16% of brands systematically track their visibility, while more than 20 vendors offer competing tools based on different query sets and scoring methods.
For marketers evaluating these platforms, the framework gives them a useful set of questions to bring into vendor conversations:
- Which of the four Ps does the tool measure?
- Is the data directional or decision-grade?
- Can the vendor provide the actual query set being used?
If a platform cannot clearly answer those questions, its score may be difficult to interpret or compare with another provider’s results.
The IAB’s companion AI attribution framework is scheduled for November 12. That could give marketers a more consistent way to connect AI visibility with actual business outcomes as the measurement category develops.
Nielsen Agreed to Buy DoubleVerify
The measurement industry faced another major shift in early September following Nielsen’s August 7 announcement that it had agreed to acquire DoubleVerify in an all-cash deal worth about $2.15 billion, or $13.60 per share. The offer represented a 30% premium over DoubleVerify’s previous closing price.
The deal would bring one of the major independent ad-verification businesses under Nielsen, combining capabilities across media measurement, verification, viewability, brand safety, and inventory-quality data. It also adds to the broader consolidation happening across the advertising measurement market.
For advertisers, the important question is what happens to DoubleVerify’s independence after the acquisition. Nielsen has said it intends to preserve the business’s independence, but we’ll have to wait and see how that plays out in real-time.
If your measurement strategy hinges on DoubleVerify, start auditing your active contracts, incoming data streams, and analytics pipelines now. To prevent operational gaps during the migration, it is also smart to establish a secondary partner for tracking brand suitability and viewability standard metrics.
Meta Agreed to $18B Settlement Across 29 States in Teen Safety Lawsuit
On August 26, Meta reached a settlement with a coalition of 29 state attorneys general over teen-safety claims involving Facebook and Instagram. All 52 state and territorial attorneys general were involved in the broader process. Reports have put the potential value at up to $18 billion, with around $5.3 billion reportedly tied to whether YouTube and TikTok adopt similar measures over the agreement’s 10-year term. Meta recorded a $10 billion legal expense in Q3.
For advertisers, the product changes matter more than the headline number.
Users aged 13–17 will face a default two-hour daily limit across Facebook and Instagram, followed by a lockout unless a parent gives permission to disable it. Meta is also introducing a default Night Mode from midnight to 6 a.m. and School Mode from 8 a.m. to 3 p.m., which mutes notifications. Teen users will also be prevented by default from seeing like and reaction counts.
YouTube Doubled the Monetization Threshold for New Partner Program Applicants
YouTube dropped some news in early September that’ll make life harder for aspiring creators. Starting February 1, 2027, anyone applying to the YouTube Partner Program will need to clear a much higher bar with either 8,000 qualified watch hours over the past 365 days, or 20 million qualified Shorts views within 90 days. To put that in perspective, that’s double the current thresholds of 4,000 hours and 10 million views. Some good news is the 1,000-subscriber requirement stays the same, and if you’re already in the program, you’re fine, as this will only hit new applicants.
For marketers, though, the more interesting story is what this does to the creator ecosystem as a whole. Push the entry bar up, and you could end up with fewer channels getting monetized in the first place. That means advertising demand gets squeezed toward more established creators who are already past the threshold.
If you run creator partnerships, expect more competition for mid-tier channels. Brands will be chasing a smaller pool of proven inventory, and that could drive prices up. Paid YouTube campaigns aren’t immune either. As the mix of eligible channels shifts, inventory availability and campaign economics for Demand Gen and Video View campaigns could feel it too and it is something worth keeping an eye on when you’re planning budgets for next year and beyond.
Instagram Rolled Out Its First Wordmark Refresh in a Decade
In August, Instagram introduced an updated wordmark that they are rolling out globally through September and beyond. It’s only a typography update rather than a complete logo redesign, with the familiar camera icon remaining unchanged, while Instagram’s typefaces have been modernized.
For brand and creative teams, the practical task is simple. Check Instagram-specific templates, cover art, partnership materials, and other assets that still use the old wordmark, then update them as the new identity becomes more widely used on the platform.
The timing also fits into a broader shift at Meta. Instagram continues to be positioned as an important platform for younger audiences, while recent product and legal changes are reducing teen exposure to Facebook. In that context, the rebrand is less about a new logo and more about reinforcing Instagram’s position within Meta’s wider strategy.
The Merchant API Deadline Actually Hit
Google officially retired the Content API for Shopping on August 18 and began replacing it with the Merchant API. From September 1, integrations still using the old API without an approved extension started experiencing intermittent HTTP 410 Gone errors. Those failures are expected to become more frequent as Google completes the shutdown.
Standard e-commerce platforms like Shopify, WooCommerce, Magento, or BigCommerce typically managed the transition automatically, keeping most merchants safe.
However, brands operating custom Merchant Center pipelines, internal ERPs, or legacy middleware face significant exposure, especially if automated product feeds started flagging errors in early September.
FTC Enforcement Kept Rhyming
Three recent FTC actions point to a broader trend in advertising compliance.
On August 27, Cox Media Group, MindSift, and 1010 Digital Works agreed to a $930,000 settlement over claims that they could target ads using conversations captured from smart devices. The FTC found that the service did not actually rely on voice data and that consumers had not consented to its use.
Doxo agreed to a $2.1 million settlement on August 17 over search ads that used billers’ names, logos, and URLs in ways that could imply an affiliation. The FTC also sued Hims & Hers in July, alongside Utah and California, over allegations involving the sharing of sensitive health information with platforms such as Meta and Snap, as well as alleged deceptive cancellation practices.
The Hims & Hers case is particularly relevant to marketing teams because the underlying setup is common: tracking pixels on pages containing sensitive information and customer lists uploaded to advertising platforms for matching. Teams should audit what data those lists contain, where pixels are firing, and whether former employees left behind tracking configurations that nobody is reviewing.
Google Also Tightened Ad Serving
Google’s Limited Ad Serving policy expanded to cover all Google Ads, with the rollout continuing through 2028. Rather than immediately disapproving ads, Google can limit impressions based on factors such as account history, policy compliance, verification status, and industry.
Google’s updated guidance also tells Search advertisers to put their domain in headline position one and be clear when referring to other brands. Affiliates, resellers, and lead-generation businesses should pay particular attention to those requirements.
Site Reputation Abuse Enforcement Changed in Europe
Google changed how it enforces its site reputation abuse policy in the European Economic Area (EEA) at the end of August.
The policy targets third-party content published on established domains mainly to benefit from the site’s existing ranking authority. Common examples include unrelated coupon, review, financial, or commercial sections run largely by outside organizations.
Since August 30, manual actions under this policy work differently for EEA searches. Outside the EEA, the affected section can still be directly suppressed. Within the EEA, Google says the section will instead be separated in its systems and gradually evaluated independently of the host site’s reputation. The change followed discussions with the European Commission.
For publishers running third-party commercial content on established domains, the core lesson hasn’t changed: editorial ownership, topical relevance, and the real purpose of the section still matter. What’s different is how Google applies the enforcement in a market that can be important for traffic and revenue.
