SEO Talks

How do you measure whether AI search is sending you anything?

You measure it with a small set of imperfect signals rather than one clean number, because no analytics view today reports AI-driven visits with full accuracy. The honest starting point is that you assemble a partial picture from referral data, Search Console patterns, and manual citation checks, then read that picture knowing it understates and misattributes some of what is happening. Anyone promising you a tidy “AI traffic” dashboard is overselling what the tools can currently do.

The first signal is referral traffic from AI tools inside your analytics. Some assistants and AI search surfaces pass a referrer when a user clicks through, so you can filter your acquisition reports for known AI domains and watch whether sessions from those sources rise over time. This captures only the clicks that actually happen and that carry a clean referrer, which is a fraction of total AI exposure, but it is real traffic you can attribute with some confidence.

The second signal lives in Search Console. When your impressions climb while clicks stay flat, that gap is a useful proxy for answers being read without a visit, which is the pattern AI summaries and zero-click results tend to produce. You cannot prove AI caused any single impression, but a widening impressions-without-clicks trend on queries you would expect AI to answer is worth tracking as a directional cue.

The third signal is manual. Periodically ask the major AI tools the questions your pages target and note whether your brand or content gets named or cited. This is slow and unsystematic, but it tells you something the logs cannot: whether you are visible in the answer itself even when no click follows. Citations there are visibility, not sessions, so treat them as a separate column from your referral numbers. All of this is worth re-checking, since AI attribution and the referrers these tools send are still shifting and any method here may improve or break within months.

To start, set up the AI-referral filter in analytics, save an impressions-versus-clicks view in Search Console for your key queries, and put a recurring reminder to spot-check citations by hand, then read each signal for what it is and accept that the gaps are part of the measurement, not a flaw you can fix.

Related posts:

  1. How do you decide what to measure before you change anything? Decide what to measure by working backward from the outcome the change is meant to produce, then lock that in before you touch anything. The method has four parts. First,......
  2. Does requesting indexing in Search Console actually speed anything up? Requesting indexing moves a page up the crawl queue, but it does not decide whether the page gets indexed. That is the split the button's reputation glosses over: it can......
  3. How long do you wait before judging whether a change helped? Wait long enough for Google to re-crawl the changed pages and for enough data to accumulate to be meaningful, which in practice usually lands somewhere in the range of a......
  4. Is the meta keywords tag truly dead, or does anything read it? For Google ranking, the meta keywords tag is dead, and it has been for well over a decade. Google announced publicly back in 2009 that it does not use the......
  5. How do you actually measure if an SEO change worked? To actually measure whether an SEO change worked, you define the metric before you touch anything, isolate the change so nothing else confounds it, compare against a baseline plus a......
  6. What does Google actually read as a “quality” signal it can measure? Google cannot measure "quality" directly, so it approximates it through measurable proxies. There is no quality meter that reads a page and returns a verdict; instead the system infers quality......
  7. Does adding an author byline change anything for a non-YMYL page? For a non-YMYL page, an author byline has limited direct ranking effect. Its real value is trust and transparency, telling readers who stands behind the content, plus a minor contribution......
  8. Does Google actually measure user satisfaction, and how would it? Google does not have a direct "satisfaction" meter, but it almost certainly infers satisfaction indirectly from aggregate behavior and quality proxies. The honest answer to both halves of the question......
  9. Why does a redirected page sometimes still show in search? A redirected page that still shows up in search usually isn't a sign the redirect failed. Google has to re-crawl the old URL and reprocess the redirect before its index......
  10. What technical issue silently kills rankings without an error in Search Console? The dangerous failures are the ones Search Console was never built to flag. Its error reports are tuned to catch broken crawling, so a page can return a clean status,......
  11. Does AI search reward the same signals as classic ranking, or different ones? Mostly the same foundation, with a different surface. AI search leans heavily on the same underlying signals classic ranking has always rewarded (authority, relevance, and trust), but it adds real......
  12. Why do images rank in image search but not help the page? Images can rank in image search yet do little for the host page because image search and web search are separate result systems with separate judgments. An image earns its......
  13. Which Search Console metric lies to you most often? Average position lies to you most often, because it is an average smeared across every query, page, and location your site appears for, and an average can move for reasons......
  14. How do you rank B2B content when search volume is tiny? When search volume is tiny, you rank B2B content by targeting the few high-intent decision-stage queries that matter and judging success by value per visit, not traffic. In B2B, a......
  15. What happens to rankings when you remove an internal link a page has relied on? When you remove an internal link a page has relied on, the page loses the share of authority and the discovery path that link was providing, and how much it......