Signals & Data
THE DEAL ENGINE doesn't guess. It reads. Every property in America leaves a paper trail — who owns it, what they paid, what they owe, whether they're behind. Public records expose most of it for free. The edge isn't access; it's knowing which signals predict a sale, at what price, and stacking them into a score.
This doc covers two things: the signal catalog (what we mine and what each thing tells us) and the data ladder (free first, paid only once a market earns it).
See also: the-deal-engine · the-intelligence-layer · the-plays · kpis-and-reports · following-the-money · Research Pipeline
The four questions
Every signal answers one of four questions. Stack the answers and a deal falls out.
MARKET signals — WHERE to hunt
You can't out-underwrite a dying market. Pick the metro first.
| Signal | What it tells us | Read |
|---|---|---|
| Job & population growth | Demand for housing is rising | Up = tailwind |
| In-migration (USPS change-of-address, Census) | People physically moving in | Net inflow = pricing power |
| Rent-to-price ratio | How much cashflow a dollar of price buys | Higher = better yield |
| Building-permit supply | New units coming online | Oversupply caps rent growth |
| Median income | Affordability ceiling & rent durability | Rising income = sticky rent |
| Path-of-progress | New transit, rezoning, planned development | Buy ahead of the wave |
PROPERTY signals — WHAT it is
Once the market's chosen, score the asset.
| Signal | What it tells us |
|---|---|
| Assessed vs market value | Tax basis vs reality — gap hints at mispricing |
| Last-sale date & price | Cost basis, hold length, likely equity |
| Beds / baths / sqft | Comparability & unit economics |
| Condition (age, permit history) | Reno scope and risk |
| Rent estimate vs price | The cashflow verdict — does it pencil? |
OWNER / MOTIVATION signals — WHO'LL sell
A great house isn't a deal until someone wants out. This is where margin lives.
| Signal | Why it predicts a sale |
|---|---|
| Absentee / out-of-state owner | Tired of remote management; emotionally detached |
| Length of ownership (long hold) | "Tired landlord" — decades in, ready to exit |
| Estimated equity position | High equity = room to discount and still walk happy |
| Portfolio owner | Trades in bulk, thinks in spreadsheets, sells rationally |
| Owner age | Probate-adjacent; estate/lifecycle transitions |
DISTRESS signals — WHO'LL sell CHEAP
Pressure prices below market. These are the discount engine.
| Signal | The pressure |
|---|---|
| Tax delinquency | Owes the county; clock is ticking |
| Pre-foreclosure / lis pendens | Lender has filed; motivated and time-boxed |
| Code violations | Fines stacking; can't or won't fix |
| Eviction filings | Landlord fatigue at a breaking point |
| Vacancy (USPS vacant flag) | No income, all carry — bleeding monthly |
| Liens | Encumbrances forcing resolution |
ALT DATA — the real edge
Anyone can pull an assessor record. The differentiated alpha is in data most operators never touch.
| Source | Edge it buys |
|---|---|
| Aerial / satellite imagery | Roof age, lot size, deferred maintenance — condition without a visit |
| Street-view imagery | Curb condition, occupancy cues, neighborhood feel at scale |
| Short-term-rental yields (AirDNA) | True income ceiling for STR-viable assets |
| Permit velocity | Where capital is actually flowing, block by block |
| Business openings / closings | Leading edge of neighborhood momentum |
The data ladder — free first, paid later
This is the cost-discipline spine of the whole engine. County-level alpha is real and free. Paid data is an amplifier you earn into — never a starting cost.
FREE / public — start here
| Source | What it gives us |
|---|---|
| County assessor + recorder | Ownership, sales history, assessed value, beds/baths/sqft |
| Tax-delinquency lists | The cleanest distress signal, published by the county |
| Court foreclosure / lis pendens filings | Pre-foreclosure pipeline, straight from the docket |
| City code-violation & building-permit open data | Distress + reno activity, often via open-data portals |
| Census / ACS | Demographics, income, in-migration |
| HUD Fair Market Rents | Free rent baseline for underwriting |
| USPS vacancy | Vacant-property flag — carry-cost distress |
| BLS | Jobs & employment trend by metro |
PAID — layer in once a market proves ROI
| Source | What it adds | Note |
|---|---|---|
| ATTOM / CoreLogic / Reonomy | Bulk, cleaned, national property data | Speed & coverage, not new signal |
| MLS feed (RESO / RETS) | Live on-market listings | Needs a license |
| PropStream / BatchLeads | Pre-built motivated-seller lists | Convenience layer over public data |
| AirDNA | STR yield data | Only where STR is the thesis |
| HouseCanary | AVM / valuation model | Sharpens, doesn't replace, our model |
| Rentometer | Rent comps | Validates the HUD baseline |
Next: the-plays turns these signals into five concrete find-X strategies — and makes the call on which one to run first.