Baidu has given Chinese webmasters a new reason to log back into its Search Resource Platform. In mid-September, the company started rolling out a webmaster tool section called AI Traffic Analysis. The official notice landed on September 16, but Baidu says the data service has not switched on yet, so most dashboards are empty.
The early signal, though, is not in the numbers. It is in the definitions.
What Baidu plans to measure
The dashboard is built around four metrics:
- AI Citations: how often a page is cited in an AI answer.
- AI Impressions: how often users see the link to a cited page.
- AI Clicks: how often users click through from a cited page.
- AI CTR: clicks divided by citations.
There is also a citation trend chart, a running count of cited pages, and tables for top keywords and top pages. Baidu sorts those tables by clicks, caps them at the top 50,000 entries, and updates data daily for the previous day, with a warning that delays are possible.
The metric definition that changes everything
AI CTR is the detail SEOs should not gloss over. Baidu calculates it as clicks divided by citations, not clicks divided by impressions. That is a meaningful break from the Search Console logic many marketers know. A page can be cited inside an AI answer while the link remains hidden in a collapsed source list, so citations and impressions can diverge. As a result, Baidu’s CTR may understate how well a visible link actually performs.
Worth doing: calculate clicks divided by impressions yourself and watch the gap between citations and impressions as a separate signal.
What the report won’t tell you
Baidu says the report covers only the “traditional search scenario” — AI answers inside regular Baidu search results. It does not appear to include standalone products such as Baidu Wenxin. That matters because China’s AI search market is fragmented across Doubao, DeepSeek, Qwen/Quark, and Kimi, each with its own ecosystem and no shared data. For some categories, those platforms may drive more AI visibility than Baidu.
The keyword table is the real prize
The metric counts are useful, but the keyword data could be more actionable. It shows which queries trigger AI answers that cite your content — demand data that is hard to get elsewhere in China. Treat Baidu’s keyword patterns as a hypothesis to test on other AI platforms, not as a rule to copy, because each platform has different citation preferences.
What to do now
Verify your site in Baidu Search Resource Platform if you have not already, so you can collect data from day one. Record a baseline once the report goes live, because early numbers matter more for trend tracking than as benchmarks. Do not expect citation visibility to boost Baidu’s version of site weight or rankings; Baidu has not said that. And keep checking server logs for crawler activity from other Chinese AI platforms, since this report will not cover them.
Baidu is now among the few search engines separating AI citations from general search performance. For international brands, that is a rare first-party window into AI visibility in China. The dashboard is empty for now, but the definitions already show that AI visibility depends heavily on which metric you choose to watch.
Source: Search Engine Journal
