Incorporate the .num_sources and .last_seen metadata fields into likelihood/match scoring that are exposed in the Person Enrichment and Person Identify endpoints. I see this as most logical (and perhaps easiest to implement) in the Identify Endpoint. See https://api.peopledatalabs.com/v5/person/identify?pretty=true&api_key=XXX&phone=19165432081 . The phone number returns 7 matches and 6/7 records have an identical match_score of 35. Record ID fBugNJFRBNLsb2-djscybg_0000 has the most sources (5) contributing to the phone match and the most recent last_seen date (3/11/2021). Once could argue that based on these metadata values, this record should have the highest match_score.
