Reproducible test data reduces coordination cost, not just randomness.
When QA reports an issue that appears only with a particular address shape, developers need the exact same names, addresses and edge fields. AddrID combines the seed with locale, field selection and a fixed reference date so matching settings return the same synthetic record.
Store the seed in test cases, bug reports and automation logs instead of copying a complete generated profile. That keeps unnecessary personal-looking data out of logs and makes scenarios easier to move between environments.
A stable reference date matters for date-derived fields. If age is calculated from today, the same seed can drift over time. The fixed reference date removes that source of nondeterminism.
A seed is a reproduction tool. It is not a password, token or cryptographically secure random value.