Overseas Brand AI Visibility Benchmark

Name: Overseas Brand AI Visibility Benchmark (saler.ai)
Creator: saler.ai
Published: 2026-09-22
License: https://saler.ai/legal/terms-of-service

Published monthly. Aggregates overseas brands' mention rate and visibility score in AI search answers, by AI engine x market. Gated by k-anonymity and client confidentiality; sub-threshold numbers are never published.

Current period: 2026-09 (real benchmark published, caliber vs2)

The panel has accumulated 6 public competitor brands, reaching the k-anonymity threshold (k≥5). The monthly median benchmark aggregated by AI engine x market is below; all cells pass k-anonymity and are computed from real snapshots, never fabricated.

What this is

The Overseas Brand AI Visibility Benchmark is saler.ai's monthly industry snapshot of how visible overseas brands are in AI search answers (ChatGPT, Perplexity, Gemini, Claude, and others). It answers two questions:

Where is the industry waterline? On average, how often are overseas brands mentioned, and what is the median visibility score.
Where does your brand stand? A single brand's visibility relative to the panel median for the same engine and market.

A numbered, method-transparent industry benchmark is among the most citable forms of content in AI answers—part of saler.ai's own "be citable by AI" strategy. But citable requires the numbers to be real, caliber-transparent, and leak no client.

Method (reuses /methodology caliber, vs2)

Sampling: brand x AI engine x date, sampled daily; raw retained 7 days, derivatives permanent.
Caliber version: vs2. The visibility-score denominator includes only observations that actually produced an AI overview (answer_source != 'none'); questions with no overview are excluded from the denominator.
Scoring: mention rate and visibility score use basis points (bp); when scored_samples = 0, visibility score is null (not computable), never shown as 0.
SoV (share of voice): denominator includes self; when self is 0, SoV is null.

Publication discipline (k-anonymity + client confidentiality)

k-anonymity threshold k≥5

Any published aggregation cell (engine x market) must be contributed by at least 5 publishable brands; otherwise the cell is suppressed (shown as "—"), never a 3- or 4-brand sample masquerading as an industry benchmark.

Client confidentiality gate (`clients.kind`)

Only kind = 'peer' (public competitor) or kind = 'client' with written public authorization may enter the benchmark; all other client brands are excluded, and the benchmark output contains no client brand name, id, or domain. This is a hard rule, guaranteed by construction.

No fabrication

This benchmark aggregates only real snapshots. When the panel is below threshold it is honestly marked "pending"—no back-filling, no estimation, no hinting at undisclosed numbers.

Current period status

2026-09

Period

Panel brands (peer / authorized)

Publishable brands

k-anonymity threshold

Published

Benchmark numbers published?

Machine-readable per-cell results: benchmark-data.json (Schema.org Dataset, includes this period's aggregation status, no brand identifiers).

This period's median benchmark by engine x market

Engine	Market	Brands	Median mention rate	Median visibility
chatgpt	en-US	6	20.83%	0.00%
gemini	en-US	6	29.17%	25.00%
perplexity	en-US	6	40.00%	20.00%

Caliber vs2; k-anonymity threshold k≥5: any cell with fewer than 5 contributing brands is suppressed, never published. A median visibility score of 0 means the panel-median brand produced no scorable AI overview (answer_source = 'none') this period—not missing data.

Related resources

Resource	Purpose
Methodology	How visibility metrics are sampled, scored, and when not computable (human-readable)
Glossary	Core GEO / SEO / AI-visibility term definitions (AI-citation friendly)
API Docs	How to connect your brand and retrieve your visibility time series
benchmark-data.json	This period's benchmark aggregation (machine-readable Dataset)