GEO RESEARCH SERIES
Weekly GEO Experiments
A recurring series of observational experiments designed to identify patterns in how AI systems cite, link, and recommend sources across real-world topics.
How this research works
Each week, we choose one focused question, run a standardized set of prompts across major AI answer systems, record the sources that appear, and code observable page characteristics using the same framework.
The goal is not to reverse-engineer proprietary ranking systems. The goal is to accumulate comparable observations over time and separate what we can document from what we can only infer.
Every study follows the same discipline: Observation → Pattern → Inference → Limitation → Next Test.
Research log
BASELINE
AI Visibility Audit: kentlundin.com
Established the starting point for the site’s own GEO visibility before further interventions.
WEEK 1
Carry-On Luggage Recommendations
Examined which sources AI systems cited when recommending carry-on luggage—and how those sources differed from the brands being recommended.
What we are looking for over time
One experiment can reveal an interesting pattern. A series can show whether that pattern repeats across query types, industries, source types, and platforms. As the dataset grows, we will look for recurring differences in evidence quality, firsthand experience, freshness, authorship, page structure, entity clarity, and topical context.
Recurring does not mean causal. Every finding will continue to distinguish what was observed from what might explain it.
