LEARN

First-Party Data First

A rigid opaque casing encloses a luminous channel, with dim particles drifting outside and unable to penetrate, symbolizing AI visibility attribution centered on owned data, with third-party tools not serving as acceptance criteria
AI visibility attribution is based on owned tracking and server logs; third-party tools are only for cross-verification, not acceptance

IN ONE SENTENCE

AI visibility attribution runs on first-party analytics and server logs; requests that do not execute JavaScript never reach front-end analytics at all.

Attribution for AI visibility runs on first-party analytics and server logs, with third-party tools for cross-checking. The reason is practical: requests that do not execute JavaScript never reach front-end analytics, so AI-side crawling is effectively invisible there.

OUR POSITION

Third-party tools are for reading order-of-magnitude and comparing competitors, not for acceptance. Hanging acceptance on data whose definition you do not control hands the interpretation to someone else.

01

What each source can answer

First-party analytics: behaviour and enquiries after the visit, limited to people who took an action.

Server logs: who fetched what and what was returned — the only first-party evidence on whether the crawl layer is healthy.

Webmaster tools: search-side impressions and clicks, on the platform's definitions.

Web analytics: sessions and channels, blind to requests that do not run JavaScript.

02

Three attribution rules

Give AI platform referrers their own channel group so they do not land in Direct / Other — under default configurations that result stays permanently invisible.

The referrer-less portion can only be estimated, and the report must say so. Do not present an estimate as a measurement.

Count the indirect path: someone who sees the brand in an answer and then searches its name or types the URL carries no AI referrer but is still a product of visibility.

03

A note on scale

Cross-site studies average around 1% of total traffic. A small base means large percentage swings, so a single month proves nothing.

The value at this layer is trend and structure rather than current absolute volume — report a trend line, not a month-on-month comparison.

Data behind this page

1.08%

AI referrals as a share of total site traffic

Source：Conductor, across 13,770 domains，2025

Sources

  1. [1]Google users are less likely to click on links when an AI summary appears．Pew Research Center．2025-07-22

Updated 2026-08-10