Predictive seller scoring
A model that estimates how likely a homeowner is to list in the next 12 months.
What it does (the operator translation)
Predictive seller scoring tools — SmartZip, Likely.ai, Catalyze AI, Offrs — score households in your farm or ZIP based on demographic, life-stage, equity, and behavioral signals. They rank a list. You work the list. Vendors claim 70%+ accuracy (SmartZip review at HousingWire, Sitejabber Offrs reviews). No independent academic study verifies that number. The vendor-claimed accuracy is the marketing — the operator number is what your phone calls actually convert.
Why a working REALTOR cares (the breakpoint)
Three gates. Below any one of them, the foundation model on your phone plus a sphere CSV beats the vendor stack. Gate one — 30+ closed sides per year. Gate two — a salaried ISA on the phone running the list. Gate three — 8+ transactions a year in a single farm ZIP. The NAR median is 10 sides, so 75% of working REALTORS sit below all three gates. The tool is staff-completion software, not a lead source.
What this is NOT (the category-flip)
Predictive seller scoring is NOT a lead source. It tells you who. Someone still has to call, mail, and door-knock for 6 to 18 months. If that someone doesn't exist on your team, the tool is a $700-a-month subscription to a list nobody actions.
Related terms
Lead scoring · Speed to lead · Predictive analytics in real estate · Seller farming
Where this comes up in The Listing Machine
The 3-gate threshold framework lives on the Lead Generation pillar. The Listing Machine teaches the foundation-model alternative that beats the vendor stack at 12 deals a year — read-time intelligence on your sphere CSV, no $700 subscription required. Karpathy's Software 3.0 framing is the technical reason.