Locale & Market Methodology: How Dabudai Measures AI Visibility by Country and Language
Jump to
AI answers change by market.
The same buyer question can return different links and sources across countries and languages. That is why Dabudai measures AI visibility per topic locale.
In Dabudai, each topic is configured with one country and one language
All prompts inside the topic inherit that locale.
We apply locale in runs by sending the selected location in provider API requests and generating prompts in the topic language.
We lock country and language after topic creation because results are aggregated.
Mixing locales would break trend comparability and produce unreliable statistics.
If you need the same topic in two countries, create two topics. This keeps measurement valid and comparable over time
How Dabudai organizes measurement (org → topics → prompts)
Dabudai uses a simple structure so results stay consistent and comparable over time.
Organization: your company workspace. It includes company-wide dashboards.
Topics: tracking units inside the organization. Each topic is one measurement dataset with its own aggregated results.
Prompts: the buyer questions tracked inside a topic.
This matters because Dabudai aggregates links, positions, and metrics across repeated runs.
A clean structure prevents mixing countries and languages inside the same dataset.
The core rule
Locale (country + language) is set at the topic level and stays fixed. All prompts in the topic inherit that locale.
What a “topic locale” is
A topic locale is the fixed (and locked) combination of:
Country (the location used for provider runs)
Language (the language used to generate and send prompts)
Every tracked prompt inside the topic runs using that same topic locale. Country is passed in the API request as the selected location, and prompts are sent in the topic language. This prevents mixed inputs in one aggregated dataset and keeps results aligned with what users in that market see in provider interfaces.
How locale is applied in provider API runs
Locale is applied at run time, not treated as a label.
Each topic uses one fixed country + language, and all prompts inherit it.
Prompt language is set automatically
Prompt language is derived from the topic language.
Prompts are generated and sent to providers in that language.
This prevents mixed-language measurement inside one topic dataset.
Location is sent in the API request
For each provider run, Dabudai includes the topic’s country/location in the API request.
This helps align outputs with market-specific contexts.
Locale settings (quick table)
Setting | Comes from | Why it matters |
Language | Topic language | Determines the language prompts are sent in |
Country / location | Topic country | Anchors runs to a specific market context |
What this means for your data
Inside one topic, all results come from:
the same country
the same language
the same topic prompt set
This keeps topic-level dashboards market-consistent and comparable over time.
Why locale cannot be changed after setup
Dabudai aggregates data across many runs within a topic dataset.
We aggregate positions, link occurrences, visible sources, and derived metrics (for example: SOV and Recommendation Rate).
If you change country or language after tracking has started, the dataset becomes mixed.
Mixed datasets create misleading trends and incorrect comparisons.
That is why topic locale is locked after creation.
What goes wrong when locales are mixed (quick table)
If you change… | You create… | Result |
Country | Mixed market outputs | Links and sources stop being comparable; SOV and rates become misleading |
Language | Mixed-language prompts and answers | Measurement drift; trend lines lose meaning |
Both | Corrupted aggregation | Metrics no longer reflect one real market over time |
How to measure the same topic across multiple countries
If you want to track the same topic in two markets, do not reuse one topic.
Create one topic per country + language, because locale is fixed and results are aggregated within a topic dataset.
The recommended setup (example)
Topic A: “AI Search Intelligence” — United States · English
Topic B: “AI Search Intelligence” — Germany · German
Both topics can use the same buyer questions.
Prompts will be sent in the topic language, and runs will use the topic country in the provider request.
Each topic produces outputs that match its market.
Goal → setup (quick table)
Goal | Correct setup |
Track one topic in two countries | Create two topics (one per locale) |
Compare performance over time | Compare within the same topic (same locale) |
Expand to a new market later | Create a new topic for that locale |
Why this improves validity
This approach keeps:
clean aggregation per market
comparable trends over time (no mixed locale data)
accurate page-level link and source lists as shown in provider answers
Video: Organization vs Topic dashboards (avoid mixed-locale data)
This video explains how to use Dabudai dashboards without mixing markets.
It shows the difference between company-wide reporting and topic-level execution.
Key rule: A Topic is one market dataset (country + language locked).
What the video covers
The dashboard hierarchy: Organization → Topics → Prompts
Why topics are locale-locked (country + language)
When to use Organization vs Topic views
How to go from “what changed” to “what to do next” (repeatable workflow)
When to use which view (quick table)
What you want to do | Best view |
Report overall trends to leadership | Organization dashboard |
Diagnose why performance changed | Topic dashboard |
Decide what to ship next | Topic dashboard |
Compare results fairly | Same Topic (same locale) |
Organization-level dashboards (company-wide)
In the Organization dashboard, you see a roll-up of performance across the company.
Best for:
leadership updates
weekly reporting
spotting overall movement in metrics
Key idea: Organization dashboards summarize multiple topics and prompt sets.
Topic-level dashboards (market-accurate)
In a Topic dashboard, you see detailed results for one measurement dataset:
that topic
all prompts inside the topic
the fixed country + language (locale)
Locale is fixed after topic creation and cannot be edited. This view is best for execution because it reflects one market setup.
Market setup checklist (copy-paste)
Use this checklist when creating a topic.
Key rule: One topic = one market dataset (country + language). Locale is locked after creation.
Topic setup checklist
Pick a topic that matches a real buyer intent area.
Choose one country where you sell or want to grow.
Choose one language your buyers use in that market (prompts will be sent in this language).
Select your providers (ChatGPT, Google AI Overviews, Google AI Mode).
Add 20 buyer questions for that topic (ICP-relevant).
Start tracking and keep the locale fixed for the measurement window (data is aggregated).
If you need another market, create a second topic with a different country/language.
Common mistakes to avoid (quick table)
Mistake | Why it breaks measurement | Better approach |
Mixing markets in one topic | Aggregates incomparable outputs | One topic per market |
Trying to change country or language mid-topic | Creates mixed datasets and unreliable trends | Locale is locked; create a new topic |
Using different prompt formats over time | Introduces prompt drift | Keep a stable format + change log |
Comparing topics with different locales | Not an apples-to-apples comparison | Compare within the same locale (or use separate topics per market) |
FAQ
Why can the same buyer question produce different results in different countries or languages?
Providers can return different links and sources by market.
Country and language affect what pages are considered most relevant.
How is locale applied in Dabudai?
Locale is applied at run time.
We send the selected country/location in provider API requests, and prompts are generated and sent in the topic language.
Why is locale set at the topic level, not per prompt?
Because Dabudai aggregates results across prompts and runs within a topic dataset.
Topic-level locale keeps the dataset consistent, and all prompts inherit that locale.
Can we change country or language after a topic is created?
No. Topic locale is locked to protect the aggregated dataset from mixed statistics.
If you need a new locale, create a new topic.
If we want two countries, can we reuse the same prompts?
Yes. You can reuse the same buyer question set across topics.
Just keep each topic tied to one fixed country + language.
Can we compare performance across countries?
Yes, but only by comparing separate topics (one topic per locale).
Avoid mixing locales inside one topic.
Which dashboard should we use for execution?
Use the topic dashboard. It reflects one market setup (fixed country + language) and one prompt set.
