Document every verifiable fact from primary sources — not from the listing you are auditing.

Compare every mapped field against your live listings. Wrong, absent, correct — marked field by field.

Run a structured question set across five buckets and grade every answer on a four-level scale.

Route each failure to the field that caused it. Attributes come before copy, and the Prompts tab gets read.

Same questions, same clean context, after the lag. The before-and-after scorecard is the deliverable.