The reporting gap your respondents named is answerable without headcount, in three lines rather than a new tool.
Citation share on a frozen prompt set. Write the twenty questions buyers ask, keep them unchanged quarter to quarter, and report the percentage of answers that name you, per engine. The frozen part is what makes it comparable: change the prompts and you have measured two different things.
AI-referred sessions as their own channel line. Some of it arrives as direct or branded search days later, so it always under-reports, but the trend is readable and it is one channel group to add in GA4.
Unlinked mentions. Named in an answer without a link appears in no referrer report, and it is the leading indicator.
All three sit next to the rankings rather than replacing them. One caveat for the report: absence is stable and presence is not, so repeat each prompt several times before calling a movement real.
Did attribution lead in both years of the survey, or did it climb?
The shift is real, but it's masking a harder problem: most teams still can't measure what actually moved the needle before attribution got complicated, so adding a better attribution model to broken baseline data doesn't fix much. The traffic loss is one symptom; the inability to connect any of it to business outcome is the actual disease.
The reporting gap your respondents named is answerable without headcount, in three lines rather than a new tool.
Citation share on a frozen prompt set. Write the twenty questions buyers ask, keep them unchanged quarter to quarter, and report the percentage of answers that name you, per engine. The frozen part is what makes it comparable: change the prompts and you have measured two different things.
AI-referred sessions as their own channel line. Some of it arrives as direct or branded search days later, so it always under-reports, but the trend is readable and it is one channel group to add in GA4.
Unlinked mentions. Named in an answer without a link appears in no referrer report, and it is the leading indicator.
All three sit next to the rankings rather than replacing them. One caveat for the report: absence is stable and presence is not, so repeat each prompt several times before calling a movement real.
Did attribution lead in both years of the survey, or did it climb?
Last year most people were worried about losing organic traffic, this year it’s attribution
The shift is real, but it's masking a harder problem: most teams still can't measure what actually moved the needle before attribution got complicated, so adding a better attribution model to broken baseline data doesn't fix much. The traffic loss is one symptom; the inability to connect any of it to business outcome is the actual disease.