Home/The Deal Engine/The Deal Engine
1,030 words · ~5 minShareable

The Deal Engine

Common Assets is the finder + connector for cashflowing real estate. The Deal Engine is what makes the finding part work: a data-intelligence system that takes messy real-estate data in and pushes ranked, underwritten, explained opportunities out the other side.

Raw real-estate data is a commodity. PropStream, ATTOM, and Reonomy all resell the same county records to anyone with a credit card. Buying that data is table stakes, not an edge.

The point: The edge isn't that we have data — it's that we have a learning machine that turns data into funded deals.

The honest edge

Three things, and only three things, separate us from a $99/mo data subscription:

  1. Better models — valuation, rent, and motivation-to-sell scoring that beats off-the-shelf accuracy in the markets we actually work.
  2. The closed loop — we don't stop at leads. We underwrite, fund, track the outcome, and feed the result back. That compounds. → following-the-money
  3. Focus — we don't try to score every property in America. We pick where to hunt, then mine hard inside that lane.
Our edge Compounds
  • Models tuned on real outcomes, not generic priors
  • Underwrite → fund → track → recalibrate
  • Every funded deal sharpens the next score
  • Specialized to our markets and our buy-box
Buying data Commodity
  • Same county records everyone else rents
  • Static — no feedback, no learning
  • Generic scores tuned for no one
  • Resold to every competitor identically
Better modelsClosed loopFocusNot just data

The two-stage funnel

The strategy has a shape. You don't analyze every house in the country — you decide where to hunt, then mine what's inside. Top-down picks the battlefield; bottom-up picks the targets.

01
Top-down market selection
Rank metros and submarkets on cashflow, growth, and supply. Choose where to hunt before touching a single property.
02
Bottom-up property & owner scoring
Inside the chosen lane, score every property and owner on value, equity, distress, and motivation to sell.
03
Instant underwriting
Run cap rate, cash-on-cash, DSCR, ARV, and IRR deterministically against kill-floors.
04
Ranked opportunities
A daily list with a verdict — pursue, watch, or pass — and the reasons behind it.
Top‑N
ranked deals out daily
2‑stage
market-first, property-second funnel
1 list
underwritten, explained, verdict-tagged

The closed loop is the moat

Every deal we underwrite and fund produces an outcome — real rent collected, real rehab spent, real sale closed. That outcome doesn't just sit in a ledger. It feeds back and recalibrates every model that touched the deal.

Six months in, our scoring is tuned on our deals in our markets. No off-the-shelf tool can replicate that, because no off-the-shelf tool gets to see whether its own predictions came true. The data engine and the capital engine train each other.

The compounding loop
Find
Models surface candidates
Underwrite
Deterministic economics
Fund
Capital deploys
Track
Real outcomes land
Recalibrate
Every model sharpens
Fig. 1 — The capital engine and the data engine train each other. Off-the-shelf tools never close this loop.

Where the edge concentrates

Not every part of this system is precious. Knowing which parts are commodity and which parts are the moat keeps the build honest and the priorities right.

The build order follows the edge: stand it up on free public data first, beat Zillow on accuracy in one market, then scale to the next lane.

The point: Anyone can buy the records. Almost no one builds the machine that turns those records into funded deals and then learns from every one.

Where to go next