What GEO methods actually hold up: 45 studies and a cold shower
The widely cited “40% visibility lift” comes with a condition: the source was already in a fixed context. It proves neither discoverability nor durable traffic.
Key numbers
- Scope: 45 studies published between November 2023 and July 2026 (plus one earlier preprint accepted at EMNLP after the window opened), together with related RAG and evaluation work.
- Core claim: GEO is not a single ranking task but a stochastic, partially observable pipeline — search activation, crawling and indexing, retrieval, reranking and context allocation, citation, prominence, factual absorption, fidelity, user behaviour.
- The most reproducible levers are topical relevance and context position; generic heuristics transfer poorly, competition erodes individual gains, and citation-oriented rewrites can impair retrieval.
- Evidence boundary: already-retrieved content can causally change whether it is cited or used, but no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability or downstream behaviour.
- Commercial audits also reveal low source overlap, substantial run-to-run variability and persistent fidelity gaps.
Method and sample
arXiv survey “Optimizing Visibility in Generative Engines: A Critical Survey of Generative Engine Optimization (2023-2026)” (2026-07-15): 45 studies selected by a defined publication window, contributing a multistage formal model, a visibility vector separating discoverability, citation, absorption and economic outcomes, an evidence hierarchy and a reproducible protocol based on repeated measurement, paraphrases, controls, human validation and multi-actor interference.
What it means for brands
Treat any “+X% visibility” claim as conditional: it was measured with the source already inside the context window. The real bottleneck is upstream — whether you can be retrieved at all, and whether you appear consistently for the same prompts. That is why distributions and trends matter more than a single run.
How SourceSight handles it
SourceSight follows the survey's advice: each audit runs a full prompt set with repeated runs, and reports mentions, position and sources per engine so “not retrieved” and “retrieved but not written into the answer” stay separate.
Limits
The survey synthesises evidence rather than adding new experiments; the reviewed studies skew by industry and language, and the conclusions are still evolving.
References
Run it on your own project
What the same prompts return across 10 engines is only knowable by running them.
Related research
Same engine, different surface, different answer: test app, web and API separately
On the same platform and the same questions, app and web cite systematically different source sets — and the API shows something else again.
Methods · 2026-10-07How to design a retest you can trust: five moves so noise does not pass as progress
AI visibility is a distribution, not a point: fix the prompts, repeat the runs, add controls, paraphrase, validate by hand — or you cannot attribute the change.