Criteria age is not the same as criteria accuracy

A lender-confirmed buy box can remain correct for six months or longer. The timestamp should tell users when it was confirmed, not silently exclude the lender. A revised guideline, portal notice, rate sheet, or restricted-industry list should trigger review.

Exactly. I would separate “last confirmed” from “status.” Criteria stay active unless corrected, replaced, or withdrawn. The system can flag an older record for rechecking without changing whether the lender appears in matching.

The practical trigger for us is a changed approval or decline that conflicts with the recorded buy box. We report the discrepancy with the source or lender response, but we should not overwrite the old record until the correction is reviewed.

Keeping the prior source matters. The audit trail should show what changed, who confirmed it, and the effective date. That gives brokers context without treating age as proof that accurate criteria became wrong.

Would a routine confirmation from the lender reset the date even if nothing changed? That would distinguish an unchanged buy box from one nobody has rechecked.

Yes. “Confirmed unchanged” is a meaningful event and should update the last-confirmed date without rewriting the source history.

I would store observed, confirmed, effective, superseded, and withdrawn as separate events. One generic “updated” timestamp loses too much context.

And a user report should open a review, not immediately hide the lender or overwrite lender-confirmed fields. Evidence still needs to be reconciled.

That protects merchants from a false negative caused by an arbitrary age cutoff while still making uncertainty visible.

The rule is clear: age prompts verification; it never becomes an automatic exclusion. Status changes only when evidence changes.