01
One-sentence definition
Not merely a word, but an object with identity, attributes and relationships.
02
What problem it addresses
Inconsistent naming across a brand, company, product or key people can create ambiguity for search and AI systems.
03
Fit and non-fit
Good fit
- Brands with multiple names across sites, social, media and knowledge platforms
- Content systems with complex product, person, place or organization relationships
Not a fit
- One-time renaming without source-data maintenance
- Brand facts that are unconfirmed or edited independently by many teams
04
Prerequisites
Canonical names, aliases and unique identifiers
Core attribute and relationship list
Fact-change and review ownership
05
Inputs and outputs
| Inputs | Outputs |
|---|---|
| Brand, product, person and organization data | Entity and alias table |
| Existing names and descriptions across platforms | Attribute and relationship map |
| Authoritative public sources | Cross-platform consistency backlog |
06
Standard steps
Enterprise knowledge graphs
Brand fact sheets
Organization structured data
07
Decision criteria
Align canonical names and aliases
Maintain key properties and relationships
Use public sources that corroborate one another
08
Common failure modes
An entity is the same as a keyword
Once the brand name is consistent, other facts need no maintenance
09
Metrics
Name consistency
Share of key touchpoints using canonical names or recorded aliases.
Attribute conflicts
Number of conflicting versions of key entity attributes.
Correction completion
Share of identified conflicts corrected and reviewed.
10
Case anatomy
Aligning brand names and product relationships
Context
Channels used legacy names, making it difficult for users and systems to determine whether they referred to the same entity.
Approach
A canonical-name, alias, attribute and relationship table was created, followed by touchpoint-by-touchpoint corrections.
Key lesson
Entity governance continuously maintains factual relationships; it is not a single brand description.
11
Tools and templates
12
Revision history
Version 2.0
2026-07-15
Upgraded to a deep knowledge unit with fit boundaries, inputs and outputs, metrics, a case and claim-level evidence.
Related terms
Evidence and sources
Princeton University and research collaborators
2023-11-16
Generative Engine Optimization
This source supports: The research framing, experimental methods and visibility discussion behind GEO.
Google Search Central
2026-07-15
Introduction to structured data
This source supports: Structured data helps machines understand entities and properties but does not guarantee search presentation.