Why this metric exists
Since the broad rollout of AI Overviews, clicks on informational queries have fallen systematically even though rankings remain stable. Position, impressions and CTR do not explain the effect, because the answer already happens in the SERP — the click becomes unnecessary for the user, the result does not become worse. AAR separates this structural effect from the normal noise of seasonality and ranking movement.
The formula
AAR = 1 − (CTR · Impr)_post / (CTR · Impr · Q_norm)_pre
The click-weighted values of a query group are compared before and after the AI Overview appears; the normalisation factor Q_norm removes seasonal and volume-driven fluctuations. The range runs from 0 (no absorption) to 1 (complete absorption). Negative values indicate click gains — typically when your own domain is cited as a source inside the AI Overview and receives additional clicks.
Use in practice
- Calculate monthly per query group: not as a domain average — absorption is strongly intent-dependent.
- Portfolio view: query groups with high AAR and no own citation are the loss zones; groups with negative AAR are the gain zones showing that citability work is paying off.
- Prioritisation: high AAR + high commercial weight = top priority for passage and entity optimisation with the goal of being cited as a source.
- Reporting: AAR explains click declines to stakeholders structurally — instead of mislabelling them as a ranking problem.
Limits
AAR is an approximation: it requires a clean pre-baseline, reacts to SERP layout changes beyond AI Overviews, and is statistically unstable for small query groups. Since August 2026, AAR analyses can be triangulated with the official impression data from the GSC generative AI report — its current data-quality limits taken into account.
Related terms
AAR quantifies the core effect of the zero-click era and complements citation rate (is your domain cited?) and CTR drift. The detailed traffic analysis is in AI Overviews & traffic absorption.
AAR separates the structural effect from ranking noise
Click losses caused by AI Overviews are not a ranking problem and cannot be measured with ranking KPIs. AAR makes the absorption effect visible per query group — and shows where citations win clicks back.
FAQ on AAR
What does an AAR of 0.4 mean? ▾
That the query group — normalised for seasonal and volume effects — loses around 40 percent of its former click potential to the AI answer. The value quantifies the structural effect of the in-SERP answer, not a ranking decline.
Can AAR be negative? ▾
Yes. Negative values indicate click gains — typically when your own domain is cited as a source in the AI Overview and the source link brings additional clicks. Negative AAR in important query groups is the success signal of citability work.
At what level is AAR calculated? ▾
Per query group, monthly. A domain average hides the differences that matter, because absorption depends strongly on intent: informational groups are absorbed heavily, transactional ones barely.
Does the GSC generative AI report replace AAR? ▾
No — it complements it. The report delivers official impressions from generative features but no click or query mapping. AAR quantifies the click effect; the report data helps with triangulation.