Paid per address for six weeks. 470 were already in the client's CRM.
Six weeks of paying someone per address and the invoice math came out worse than if I had done nothing at all.
Setup. I hold a small portfolio and started running leads on the side for two investor clients, both buying 1970s small multifamily and tired single family in my metro. I hired a part-timer at $3 per delivered address, each one with a photo and a two line condition note. Over six weeks she brought me 1,180 addresses. That is $3,540 in driver pay, plus $190 for a skip trace batch on the 300 that looked worst.
Client A's acquisitions manager ran my list against their CRM and came back with 470 already in their system, a chunk of them from mailers they had sent twice. Another 260 got dropped for being outside their box on unit count or year built. Of what was left they pursued 14 and closed 1. My fee on that close was $2,000.
$3,730 out, $2,000 in, six weeks of my evenings managing a route.
What went wrong, in order. I paid on delivery instead of on acceptance, so a duplicate cost me exactly what a live lead cost me. I never asked either client for a suppression file, and when I finally did, both had one sitting there and sent it in an hour. Third, I wrote the driver's criteria around visible distress instead of the client's actual buy box, so she shot every house with tall grass regardless of size or vintage.
What I would do differently. Get the suppression file before day one and load it on the driver's phone. Pay a higher rate per accepted address and nothing for a rejected one, with the client sending accept or reject inside 72 hours so the driver still gets paid on a normal cycle.