Index — Full Listing
09 Entries · CC BY-SA 4.0
The
Index
The lab notebook of JUN & GEO, in three files. Failures are the post-mortems — what broke, what it cost, what changed. Milestones are the records of what got built: the framework, the scoring system, the method. Research is the evidence base — published findings credited to their authors, our own measurements labeled as ours. Every entry is published in full; nothing here is a summary of a summary.
White Paper — 旗舰V1.0 · 2026-10-20
WP Measured, Not Guessed: A Field Methodology for AI Visibility (GEO)The full methodology in one document: two-part measurement, three rates, three-sample discipline, the cognition audit, fix operations, and the reporting standard. CC BY-SA 4.0. White Paper 2026-10-20Failures — 翻车记录03 PUBLISHED
F.01 What AI Thinks Your Factory MakesThree AI systems described one Cixi appliance OEM three different wrong ways. Mention 100%, citation 0%. Failures 2026-10-20 · Published F.02 Deleted in Five Days: Our Wikidata Item Didn't SurviveCreated October 12, deleted October 17 under the notability policy. A post-mortem against the policy's three criteria. Failures 2026-10-20 · Published F.03 We Scored AI Answers From a Single Sample for Four VersionsFour versions of our scoring system graded brands on single draws. What it cost, how we caught it, what v5 changed. Failures 2026-10-20 · PublishedMilestones — 达成记录03 PUBLISHED
M.01 The Two-Part GEO Audit: Exposure × CitationExposure (awareness, share of voice, sentiment) × Citation Probability (citation rate, presentation quality), on a 2×2 grid. Milestones 2026-10-20 · Published M.02 From v1 to v5: Building an AI Visibility Scoring SystemA citation field that stored null, a third angle nobody measured, and a sampling fix. The build log, v1 to v5. Milestones 2026-10-20 · Published M.03 Five Minds, One PracticeKing on mechanism, Indig on measurement, Solis on execution, Fishkin on zero-click, Ray on evidence — disagreements kept. Milestones 2026-10-20 · PublishedResearch — 技术研究03 PUBLISHED
R.01 Cited Is Not RecommendedLily Ray's retests: cited vendors excluded from the recommendation in 69% of cases (Jun 2026), 83% by Sep 2026. Research 2026-10-20 · Published R.02 One Sample Lies: Why We Run Every Prompt 3 TimesKevin Indig: the same prompt run three times, consistent citations 2.2% of the time. Our sampling standard. Research 2026-10-20 · Published R.03 llms.txt: Necessary, Not SufficientOur measurement: near-zero direct effect on citation rates. Hygiene, yes. Ranking lever, no. Research 2026-10-20 · PublishedLicense. All articles in this index are licensed CC BY-SA 4.0 (Creative Commons Attribution-ShareAlike 4.0 International). Evidence labels: figures credited to named researchers are published research; figures marked as ours are from this practice's own measurement work.