Locale variance tracking across AI answer engines
Locale variance is the difference in AI answers for the same prompt across markets. Collection runs through residential egress in the target country with the language set explicitly, so a UK number and a German number are genuinely comparable.
How it differs by engine
Collected by running the same prompt through per-market egress and locale parameters.
Google AI Overviews locale variance
Collected by repeating the prompt per market with country and language parameters.
google_ai_overviews
Google AI Mode locale variance
Repeated collection per market with country egress and language set.
google_ai_mode
ChatGPT locale variance
Controlled by prompt language, reinforced by egress country.
chatgpt
Perplexity locale variance
Controlled by prompt language plus residential egress in the target country.
perplexity
Microsoft Copilot locale variance
Controlled by egress country and language, following Bing's market handling.
copilot
Gemini app locale variance
Controlled by account region and egress country together.
gemini
DeepSeek locale variance
Controlled by prompt language and egress, with language the stronger variable.
deepseek
Qwen locale variance
Controlled by prompt language and egress country.
qwen
Doubao locale variance
Controlled by domestic egress and account provenance.
doubao
Kimi locale variance
Controlled by prompt language, with document-language availability the dominant factor.
kimi
Baidu ERNIE locale variance
Controlled by domestic egress and Baidu's own regional parameters.
baidu_ernie
Sampling
What the API returns
{
"engine": "google_ai_overviews",
"field": "locale_variance",
"prompt": "best ai visibility tracking tool",
"country": "GB",
"language": "en",
"n": 30,
"locales": [
{
"country": "GB",
"language": "en",
"share": 0.34,
"divergence": 0.02,
"unique_domains": 4
},
{
"country": "DE",
"language": "de",
"share": 0.19,
"divergence": 0.15,
"unique_domains": 11
}
]
}Response fields
| Key | Type | Notes |
|---|---|---|
| country | string | ISO 3166-1 alpha-2 country of the egress used. |
| language | string | ISO 639-1 language the prompt was issued in. |
| share | float | Brand share within that locale. |
| divergence | float | Distance from the global mean for this prompt. |
| unique_domains | integer | Domains cited in this locale and nowhere else. |
Locale variance FAQ
- What is locale variance tracking?
- Locale variance is the difference in AI answers for the same prompt across markets. Collection runs through residential egress in the target country with the language set explicitly, so a UK number and a German number are genuinely comparable.
- How is locale variance sampled?
- Each locale is its own sample set. A cross-market comparison at n=30 per market costs what it says on the tin, which is why we surface the arithmetic before you commit.
- What does locale variance cost?
- Each locale multiplies the collection count. Two markets at n=30 is 60 credits per prompt.
Get locale variance on your prompt set
Track it in the dashboard or pull it in through the API. Send us the prompts, the brands and the markets, and we'll come back with a per-engine coverage plan and the credit arithmetic behind it.