Why discovery, retrieval, citation, absorption, prominence, action, and conversion must be measured as separate stages.
Visibility is a sequence, not a position
The critical survey of GEO models generative visibility as a stochastic, partly observable pipeline. A practical version is: discoverability → retrieval → citation selection → citation absorption → prominence → brand visibility → user action → conversion.
Discoverability asks whether the engine can access and recognize the information. Retrieval asks whether it enters the candidate source set. Citation selection asks whether the system chooses it for attribution. Absorption asks how much the source contributes to facts, language, evidence, or structure. Prominence asks where and how strongly it appears. The later stages track brand representation, behavior, and economic value.
Citation selection is not citation absorption
Zhang, He, and Yao separate two outcomes that are often collapsed. Citation selection measures whether an engine chooses a source. Citation absorption measures the depth of the source’s contribution to the answer.
In their cross-platform dataset, citation breadth and citation depth diverged: systems that displayed more sources did not necessarily show greater average influence from each fetched page. This does not prove that one platform is better. It shows that a citation count alone cannot tell us how much a source shaped an answer.
A source can be visible as a link but nearly invisible in the language of the response.
Measure the failure stage before changing content
If a brand never appears, the problem may be discovery, retrieval, entity evidence, or category relevance. If the brand appears but its own pages are never cited, the issue may concern the source ecosystem rather than basic awareness. If a page is cited but the product is not recommended, the answer may lack compelling comparative evidence—or independent sources may favor competitors.
A useful measurement program records the exact prompt, platform, date, search activation, brands mentioned, first brand, citations, cited domains, own-domain presence, approximate position, representation accuracy, and downstream activity. Repeated runs and prompt paraphrases help expose instability.
The goal is diagnosis. Optimizing the wrong stage can produce activity without progress.
Questions about this topic
What is citation absorption?+
Citation absorption is the degree to which a selected source contributes facts, wording, evidence, or structure to the generated answer.
Why is citation count insufficient?+
Two sources can both be cited while one supports several important passages and the other appears only as an additional link. Their practical visibility is different.
Can one AI Visibility score represent the whole pipeline?+
A composite score can summarize performance, but it can hide the failing stage. Keep the underlying measures visible so teams can distinguish discovery, citation, representation, and business outcomes.
References
Sources are listed in APA 7 style. Preprints are identified as such and should not be treated as peer-reviewed findings unless separately published.
- Martinez, O. (2026). Optimizing visibility in generative engines: A critical survey of generative engine optimization (2023–2026) [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2607.14035
- Zhang, K., He, X., & Yao, J. (2026). From citation selection to citation absorption: A measurement framework for generative engine optimization across AI search platforms [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2604.25707
- Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., & Narasimhan, K. (2024). GEO: Generative engine optimization. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (pp. 5–16). Association for Computing Machinery. https://doi.org/10.1145/3637528.3671900