Playbook: a 1-hour fraud audit before you scale a pop campaign
Before you 5x spend, run the smell test:
— Click-to-impression ratio per zone. Pop should sit low and stable; a zone at 2x the network median is firing twice.
— Datacenter IP share. Over 8% non-residential on a pop source you're being fed proxy junk.
— Time-to-conversion histogram. A wall of sub-3-second conversions = pre-filled or injected.
— Unique-IP / click ratio. Below 0.6 means the same boxes hammering you.
— Device-fingerprint entropy: low diversity on a 'global' zone is a tell.
Any two flags red? Cap, don't scale.
— confidence: high
Pop Wire
@pop_wire