The rule that makes it work
Identity stitching works only when hashing is consistent (same normalisation, same algorithm) and the field is marked as an identity with the correct namespace, such as Email_LC_SHA256.
Add hashed email as a secondary identity
- Open the schema linked to the dataset (not the dataset itself).
- Add a field for the hashed email, for example
person.email_lc_sha256. - Mark it as an identity with the Email_LC_SHA256 namespace.
- Keep it secondary — do not set it as primary.
- Ingest new records — new data now stitches on the secondary identity.
Secondary-only datasets
A dataset with only secondary identities still participates in stitching, but cannot anchor a profile on its own; a primary identity must come from another dataset.
Frequently asked questions
Do I need to normalise email before hashing?
Yes. Lowercase and trim the email before hashing in every source, or the hashes will not match.
Can hashed email be the primary identity?
It can, but many designs keep a CRM ID as primary and hashed email as a linking identity.
Last updated: