Studying ChatGPT citations in Spanish

A ranked domain list needs a clear counting rule. A page cited several times in one answer can otherwise outweigh a domain appearing in many different answers. This study design uses one domain occurrence per answer and preserves the cited URLs for inspection.

Status: Study design
Language: Spanish
Results: No dataset published

A protocol you can reuse

Proposed study design. No collection has been run and no results are implied.

Which domains does ChatGPT cite across different Spanish-language question types?

Choose 30 Spanish prompts: ten category recommendations, ten product comparisons, and ten factual explanations. Keep a single market and fixed account context. Collect each prompt on three separate days within one week: 90 planned answers.

Extract every cited URL, normalize only its host for domain counts, and count a domain once per answer. Keep the original URL, redirect target when checked, and passage association. Label failed collection separately from an answer with no citations.

Report answers citing each domain divided by collected answers, then break the result down by the three question groups. A broad publisher leading one group may not lead another. Publish the prompt list and collection coverage with the result.

Collection record

Download text

STUDY RECORD — EMPTY TEMPLATE
study_id: [assigned identifier]
prompt_id and exact wording: [text]
question group: [category]
language and market: [settings]
platform: [approved platform name]
collection date and account context: [details]
collection state: completed / failed
AI Overview present (Google only): yes / no / not applicable
answer reference and preserved text: [reference]
brand presence decisions: [brand, yes/no/ambiguous, passage]
cited URLs and supporting passages: [references]
reviewer and exclusion reason: [details]

What the study can distinguish

Distinguish a domain cited once in many answers from a domain cited repeatedly in one answer.
Break the result down by question type so a broad publisher does not obscure a specialist source.
Keep owned sites, competitor sites, and independent publishers visible as separate categories.

Collection and interpretation

Choose a fixed Spanish-language prompt set and record the collection window.
Extract the URLs actually cited in each answer. Normalize hostnames, preserving original URLs and subdomains where they identify distinct publishers.
Count answers citing each domain, show the denominator, and inspect the pages behind the leading results.

ChatGPT visibility tracker

Track how ChatGPT mentions, cites, and frames your brand across prompt sets.

Free AI visibility report

Get a live baseline for your brand across assistants, citations, and competitors.

AI visibility research

Benchmark pages and first-party findings about assistant answers and source behavior.