Structuring verification results in your own database
APIZONE
·
·updated 8 Sep 2026 ·
1 min read
How you store a verification result determines how easily you can audit it, reuse it, and delete it later. A schema that ages well.
On this page
Separate the raw from the decision
Keep the raw response (or a reference to it) in one place, and your derived decision (matched, not matched, could-not-verify, plus your similarity score) in another. You will query the decision constantly and the raw rarely.
Key by identifier and purpose
Index on the identifier checked and the purpose. Every check on an identifier and every check for a purpose in a period are the two queries you will run under pressure.
Timestamp and version everything
Store when the check ran and which version of your matching logic produced the decision. When you tighten the threshold, you want to know which historical decisions used the old one.