Deterministic matching

Audience: Platform admin, data or analytics engineer
Prerequisites: IDR overview →, Setup steps →, Prepare your data →, Golden Record →

Use deterministic matching when your data includes stable, specific identifiers that can be matched exactly.

What is deterministic matching?

Deterministic matching links records when one or more identifier values match exactly—like email == email or user_id == user_id. It doesn’t use fuzzy logic or similarity scoring. Matches either qualify or don’t.

For example:

john.doe@gmail.com matches john.doe@gmail.com
It does not match johndoe@company.com or john.doe+promo@gmail.com

When to use it

Use deterministic matching if:

You have structured, consistent data
You need precise control (e.g. for compliance or operations)
You want clean merges with no ambiguity
Your systems use shared, unique IDs like user ID or email

Benefits

High precision with low risk of incorrect merges
Simple to understand and debug
Fast and scalable for large datasets

Limitations

Doesn’t handle typos or inconsistent formatting
Can’t link identities across disconnected systems without shared IDs
Won’t catch cases like john.doe@gmail.com vs johndoe@company.com

How to enable it

Deterministic matching in Hightouch links records when shared identifiers—like email, user_id, or anonymous_id—have an exact match across models.

Deterministic matching is enabled automatically when you assign identifiers to your models. There’s no toggle or confidence threshold to manage.

To use it effectively, you’ll need to:

Assign identifiers in your source models
Prioritize which identifiers Hightouch should use first when evaluating matches
Optionally, set limit rules to prevent over-matching (e.g. when a shared device ID connects too many users)

Once configured, Hightouch creates identity groups by linking records that share the same value for a prioritized identifier.

Learn more → Prepare your data to build a customer identity graph