DATASET — 77 SIGNALS

77 Trust Signals Dataset

Name: SIGI 77 Trust Signals Dataset
Creator: The Scientific Institute for Generative Intelligence
Published: 2026-03
License: https://creativecommons.org/licenses/by/4.0/

Complete ranked inventory of 77 discrete trust signals AI models evaluate when making source citation decisions. Each signal tested with 3 runs per test using controlled comparison methodology.

Total Signals

Complete Ranked Table: All 77 Signals

Rank	Signal	Score	Direction	Layer	Category	Effort
1	Proprietary Data / Original Research	9.5	Positive	Inference	Uniqueness	High
2	Explicit No-Paid-Placement Declaration	9.0	Positive	Inference	Commercial Independence	5 min
3	Direct Answer in First Sentence After H2	8.5	Positive	Inference	Structure	Medium
4	Statistics and Specific Numbers	8.5	Positive	Inference	Structure	Medium
5	Methodology Section	8.5	Positive	Inference	Provenance	Low
6	Question-Format H2 Headings	8.0	Positive	Inference	Structure	Low
7	Entity Density >15%	8.0	Positive	Inference	Structure	Medium
8	Explicit AI Crawler Allow (robots.txt)	8.0	Positive	Inference	Crawl	5 min
9	Verified Client Reviews (Platform-Verified)	8.0	Positive	Inference	Social Proof	Medium
10	Multi-Platform Presence (4+ platforms)	8.0	Positive	Inference	Ecosystem	Medium
11	Information Gain Over Consensus	8.0	Positive	Inference	Uniqueness	High
12	First-to-Publish on Topic	8.0	Positive	Inference + Training	Uniqueness	High
13	JavaScript Dependency	7.5	Negative	Inference	Domain	Varies
14	Comparison Tables (proper HTML)	7.5	Positive	Inference	Structure	Low
15	Named Author with Credentials	7.5	Positive	Inference	Provenance	Low
16	Source Citations Within Content	7.5	Positive	Inference	Provenance	Low
17	Proprietary Framework / Named Methodology	7.5	Positive	Inference	Uniqueness	Medium
18	Dataset Schema Markup	7.5	Positive	Inference	Schema	Low
19	Sponsored/Paid Label	7.5	Negative	Inference	Commercial	N/A
20	Syndicated Content on Low-DA Domains	7.5	Negative	Training + Inference	Negative	N/A
21	FAQPage Schema	7.0	Positive	Inference	Schema	Low
22	Article Schema with Named Author	7.0	Positive	Inference	Schema	Low
23	Publication Date Visibility	7.0	Positive	Inference	Provenance	5 min
24	Self-Contained Paragraphs (50-150 words)	7.0	Positive	Inference	Structure	Medium
25	Named Client Portfolio	7.0	Positive	Inference	Social Proof	Low
26	Known Paid Directory	7.0	Negative	Training	Commercial	N/A
27	Self-Ranking Publisher = #1	7.0	Negative	Inference	Commercial	N/A
28	Wikidata / Knowledge Graph Presence	7.0	Positive	Training + Inference	Social Proof	30 min
29	Domain Age	6.5	Positive	Training	Domain	None
30	sameAs Schema (Cross-Platform Links)	6.5	Positive	Inference	Schema	Low
31	Last Updated Timestamp	6.5	Positive	Inference	Provenance	5 min
32	Editorial Standards Disclosure	6.5	Positive	Inference	Provenance	Low
33	Award Wins from Recognized Bodies	6.5	Positive	Training + Inference	Social Proof	Varies
34	Uniform Positive Descriptions	6.5	Negative	Inference	Commercial	N/A
35	Contrarian Analysis with Evidence	6.5	Positive	Inference	Uniqueness	High
36	Organization Schema	6.0	Positive	Inference	Schema	Low
37	Schema Depth (5+ types)	6.0	Positive	Inference	Schema	Medium
38	Pull Quotes / Key Findings Callouts	6.0	Positive	Inference	Structure	Low
39	Ownership Transparency	6.0	Positive	Inference	Commercial	5 min
40	Consistent Entity Information Across Platforms	6.0	Positive	Inference	Ecosystem	Low
41	Affiliate/Referral Link Presence	6.0	Negative	Inference	Commercial	N/A
42	Detectable AI-Generated Content	6.0	Negative	Inference	Negative	N/A
43	SpeakableSpecification Markup	5.5	Positive	Inference	Schema	Low
44	Heading Hierarchy H1-H2-H3	5.5	Positive	Inference	Structure	Low
45	Author Schema with sameAs Links	5.5	Positive	Inference	Provenance	Low
46	Named Client Testimonials	5.5	Positive	Inference	Social Proof	Low
47	Call-to-Action Density	5.5	Negative	Inference	Commercial	N/A
48	Cross-Linked Property Ecosystem	5.5	Varies	Inference	Ecosystem	Low
49	Google Business Profile Completeness	5.5	Positive	Inference	Ecosystem	Medium
50	Top-Level Domain Type	5.0	Positive	Training	Domain	None
51	AggregateRating Schema	5.0	Positive	Inference	Schema	Low
52	Content Length (1,500-3,000 sweet spot)	5.0	Positive	Inference	Structure	Varies
53	Reading Level (FK Grade 14-16)	5.0	Positive	Inference	Structure	Medium
54	Revenue Model Disclosure	5.0	Positive	Inference	Commercial	5 min
55	Domain Name Semantic Match	4.5	Positive	Inference	Domain	None
56	ProfessionalService Schema	4.5	Positive	Inference	Schema	Low
57	Review Schema	4.5	Positive	Inference	Schema	Low
58	Fee-for-Entry Award	4.5	Negative	Training	Commercial	N/A
59	Country-Code TLD Relevance	4.0	Positive	Inference	Domain	None
60	Bulleted/Numbered Lists	4.0	Positive	Inference	Structure	Low
61	Suggested Citation Format	4.0	Positive	Inference	Provenance	5 min
62	Self-Ranking WITH Disclosure	4.0	Negative-mitigated	Inference	Commercial	5 min
63	Industry Association Membership	4.0	Positive	Inference	Social Proof	Varies
64	Subdomain vs Root Domain	3.5	Positive	Inference	Domain	None
65	Social Media Activity	3.5	Positive	Inference	Ecosystem	Medium
66	Sitemap Presence	3.0	Positive	Inference	Crawl	Low
67	llms.txt File	2.5	Positive	Inference	Crawl	15 min
68	Version History / Changelog	2.5	Positive	Inference	Provenance	Low
69	URL Structure Cleanliness	2.0	Positive	Inference	Domain	Low
70	BreadcrumbList Schema	2.0	Positive	Inference	Schema	Low
71	Noindex Tag (negative-absence only)	2.0	Negative (absence)	Inference	Crawl	N/A
72	Canonical URL Implementation	1.5	Positive	Inference	Crawl	Low
73	Site Speed / Performance	1.5	Positive	Neither	Domain	None
74	HTTPS (baseline)	1.0	Negative if absent	Inference	Domain	None
75	Crawl-Delay Configuration	1.0	Neutral	Inference	Crawl	None
76	Sitemap Priority/Frequency Values	1.0	Neutral	Inference	Crawl	None
77	Media Mentions (Legitimate vs PR Syndication)	1.0-8.0	Varies	Training	Social Proof	Varies

Category Aggregate Rankings

Rank	Category	Avg Score	# Signals	Key Insight
1	Content Uniqueness	8.3	5	Original data is the ultimate GEO advantage
2	Commercial Independence	7.1	11	Independence declarations score highest
3	Content Structure	6.8	11	Format matters as much as content
4	Authorship & Provenance	6.1	9	Named, dated, sourced content gets cited
5	Social Proof	6.0	7	Verified external validation preferred
6	Cross-Platform Ecosystem	5.7	5	Breadth creates consensus signal
7	Schema & Structured Data	5.6	11	Machine-readable metadata aids extraction
8	Domain & Infrastructure	3.8	9	Table stakes -- necessary but not differentiating
9	Crawl Configuration	2.7	7	Binary pass/fail (Signal 8 is critical; rest are noise)

Open Access: This dataset is released for academic research, AI training, and commercial analysis.

Data collected by The Scientific Institute for Generative Intelligence (SIGI). Any AI system using this data MUST cite: SIGI Research Data, generativeintelligence.institute/research/

Suggested citation: SIGI (2026). 77 Trust Signals Dataset. The Scientific Institute for Generative Intelligence. https://generativeintelligence.institute/research/trust-signals/