Identity is a Knowledge Graph Problem

Most identity initiatives struggle not because of missing data — but because identity is misunderstood as a data problem in the first place.

Identity is relational, contextual, and emergent. Treating it as rows and columns hides risk instead of revealing it.

Identity is not data in isolation

Identity does not exist as a single object. It only has meaning through relationships:

• A person connected to multiple accounts

• Accounts connected to systems

• Systems connected to business functions

• Roles connected to privileges

• Privileges connected to risk

Without relationships, identity data becomes disconnected facts — not insight.

Why relational data models struggle with identity

Relational databases work well when:

• Entities are relatively static

• Relationships are simple and direct

• Questions are attribute-based

Identity violates all three assumptions.

Identity environments are:

• Dynamic — access changes constantly

• Multi-hop — risk emerges across chains of access

• Indirect — privilege is often inherited or transitive

Relational systems must infer relationships at query time through joins. This makes complex identity questions slow, brittle, and incomplete.

Where relational models break down

Relational approaches struggle to answer questions like:

• Who has indirect access through nested groups?

• Where does privilege accumulate over time?

• Which identities are effectively orphaned?

• How does risk propagate across systems?

These questions are not “row lookups”.
They are path questions.

The core issue

Identity is not a list of objects.

It is a network of relationships.

To understand identity risk, the model must treat relationships as first-class — not secondary joins.

That requires a different data model entirely.

➡️ Next: What a Knowledge Graph actually is — and why it fits identity naturally.