Nội dung
- “AI visibility” metrics are not as stable as many people think
- Google Ads quietly changes how product categories are assigned
- AI Performance reports in Search Console can be viewed with a URL trick
- Google says canonicalization may take up to two weeks to stabilize
- ChatGPT Ads now has a new Overview tab and expands to Japan and Korea
- What this means for the Vietnamese market
- References
Last week, the SEO and search marketing landscape continued shifting toward a multi-platform direction: from the stability of AI visibility data, to how Google automatically categorizes products, to new signals in Search Console and the expansion of ChatGPT advertising. For Vietnamese marketers, these are not just tech news but changes that could directly affect measurement, content optimization, and budget allocation.
-
Key points:
- “AI visibility” data can fluctuate sharply, so a single measurement should not be treated as absolute truth.
- Google Ads is beginning to emphasize self-assigned product categorization using an evolving taxonomy.
- There is a way to access AI performance reports in Search Console for platform properties through a temporary URL trick.
- Google says canonicalization issues may take up to two weeks to fully resolve; ChatGPT Ads is also expanding into Japan and South Korea.
“AI visibility” metrics are not as stable as many people think
New research cited by Search Engine Journal shows that metrics tracking visibility in AI-generated results do not reflect a fixed state. When the same query is repeatedly sent to systems such as SearchGPT, Gemini, or Perplexity, results and citation sources can change from one query to the next.
The notable point is that differences between two brands on a dashboard do not necessarily indicate a real advantage. Some of it may come from the model’s natural variability rather than from superior SEO performance or content quality from a competitor. In other words, AI visibility data should be viewed as a snapshot of a moving target, not a fixed “truth.”
For SEO and content teams, this suggests a more cautious way of reading data: prioritize trends over time, repeat measurements, and combine them with qualitative signals instead of making decisions based on a single check.
Source: Search Engine Journal.
Google Ads quietly changes how product categories are assigned
Google Ads has updated its help documentation for “product category insights,” stating clearly that the system will automatically categorize products based on an evolving taxonomy. This was already familiar to Merchant Center users, but it is now documented more clearly within the Google Ads ecosystem.
In practice, this change shows that Google continues to push product data standardization at the platform level. For retailers and performance marketing teams, that means a company’s internal categories may no longer be the only classification layer; Google will interpret products according to its own logic for reporting and optimization.
What marketers should do is check consistency between the feed, website categories, and product group structure in campaigns. As Google’s taxonomy changes over time, category reports and performance insights may shift as well.
Source: Search Engine Roundtable.
AI Performance reports in Search Console can be viewed with a URL trick
A discovery shared by the SEO community is that for platform properties in Search Console, users can add “/ai” to the path after “search-analytics” or “discover” to open AI/generative content performance reports. This trick currently works only for platform properties and does not apply to all standard property types.
Although this is not an official feature widely promoted by Google, it shows Search Console is moving closer to displaying performance signals for social and video content in greater detail. For brands investing heavily in YouTube, social video, or creator content, this could be an early reference point for evaluating which content Google is recognizing.
However, because this is a temporary access method, marketers should not build rigid reporting processes around it. It is better to treat it as an additional observation channel, alongside analytics, social insights, and conversion reports.
Source: Search Engine Roundtable.
Google says canonicalization may take up to two weeks to stabilize
In a recent exchange, Google said issues related to canonicalization may take up to two weeks to be fully processed or stabilized. This is important for websites that have recently changed URL structures, migrated platforms, consolidated content, or edited canonical tags.
From a technical SEO perspective, this statement reminds operations teams not to expect Google to respond instantly after every optimization. If a page is not selected as the canonical version right away, that is not necessarily a serious error; in many cases, the system simply needs more time to crawl, compare, and reassess.
For businesses, the safe approach is to monitor logs, index status, internal linking signals, and canonical tags consistently for at least a few days to two weeks after major changes.
Source: Search Engine Roundtable.
ChatGPT Ads now has a new Overview tab and expands to Japan and Korea
OpenAI continues expanding the advertising infrastructure for ChatGPT Ads by adding a new Overview tab and bringing the service to Japan and South Korea. This signals that the advertising ecosystem around conversational AI tools is gradually moving into a broader commercialization phase.
For marketers, this market expansion shows that the competition for visibility within AI platforms is no longer just an experiment in a few countries. As these platforms expand availability, multinational brands will need to reassess content strategy, targeting, and performance measurement by market.
From a broader SEO perspective, this trend also reinforces a reality: search behavior is fragmenting across traditional search, direct AI answers, and new advertising platforms. Optimized content therefore needs to serve citation potential, conversion potential, and visibility across multiple contexts at the same time.
Source: Search Engine Roundtable.
What this means for the Vietnamese market
For Vietnamese businesses, these five updates send a very clear message: SEO and performance measurement are entering a “multi-signal” phase, where data from Google, AI tools, and advertising platforms may not align from the start. Instead of looking for a single number to make decisions, marketing teams should prioritize repeated data over time, cross-check multiple sources, and build a response process that is slower but more reliable.
In e-commerce, Google’s automatic taxonomy assignment further highlights the importance of clean feeds, consistent categories, and clear product descriptions. In content and SEO, the instability of AI visibility and the delay in canonicalization remind us that investing in site structure, internal links, and in-depth content remains a more durable foundation than chasing short-term fluctuations.
If your business is doing SEO, content, advertising, and social video at the same time, now is a good time to review your measurement system: which pages need index priority, which sources should be monitored weekly, and which data should be used only as reference signals rather than as the sole KPI.
See more marketing news and guides at https://marketing365.vn.
Follow more updates from Marketing365 to stay on top of the latest marketing trends.
Read more articles in the same category Digital Trends.
This article focuses on the latest SEO news with a perspective for the Vietnamese market.
References
- AI Visibility Rankings Aren’t Stable – New Research Shows It’s Mostly Statistical Noise
- Google Automatically Assigns Product Categories With Evolving Taxonomy
- Hack: Google AI Performance Reports For Platform Properties
- Google: Canonicalization Issues May Take Up To Two Weeks To Resolve
- ChatGPT Ads New Overview Tab, Expands To Japan & South Korea
- Search Engine Journal — AI Visibility Rankings Aren’t Stable – New Research Shows It’s Mostly Statistical Noise via @sejournal, @MattGSouthern
