Built for a Bay Area multifamily investor
How to Find Off-Market Properties: Inside a Multifamily Signal Engine
The reliable way to find off-market properties is to watch the public records that precede a sale: ownership changes, liens, and financial distress signals sitting in county recorder data. We built a Bay Area multifamily investor a system that ingests county recorder records nightly, scores every property on 14 distress and opportunity signals, ranks it, and attaches comps, surfacing 1,800+ qualified leads a month from three live counties, re-ranked overnight.
Clients are anonymized by agreement. The system mechanics and first-party numbers below are real.
- 1,800+
- qualified leads per month
- 14
- signals scored per property
- overnight
- pipeline refresh
- 3
- counties live, built to scale to 9
Walkthrough
Watch the system run end to end
Demo shown with sample data and a fictional deal.
01 · The problem
The challenge
The investor competes for multifamily assets that almost never reach a listing. The signals that precede an off-market sale sit buried across county recorder sites, each with its own format, and surfacing them meant analysts copying records into spreadsheets by hand. By the time a property was researched, scored, and matched to an owner, the window to reach out had usually closed. Coverage was shallow, the data went stale within days, and nobody could say which leads were worth a call.
02 · The architecture
The system
County-record ingestion
Automated nightly ingestion across county recorder sources, with per-source toggles so coverage expands county by county. Three counties were live at handover, with the architecture built to scale to nine.
Distress and opportunity signals
Raw records are normalized and de-duplicated, then scanned for the 14 ownership and financial signals that tend to precede an off-market sale.
Scoring and ranking
Every property is scored and ranked so the highest-intent opportunities surface first, instead of analysts reading through undifferentiated lists.
Comps and geocoding
Each lead is enriched with comparable sales and precise geocoding, so a ranked, contactable list is ready the moment the investor opens the dashboard.
03 · The workflow
How it runs
- Ingest county records nightly
- Normalize + dedupe
- Flag distress signals
- Score + rank
- Attach comps + geocode
- Ranked call list
04 · The outcomes
Results
| Dimension | Before | After |
|---|---|---|
| Lead pipeline | scattered spreadsheets | 1,800+ qualified leads a month |
| Pipeline freshness | stale within days | overnight refresh |
| County coverage | manual, ad hoc | 3 counties scanned nightly |
| Deal readiness | manual comp + owner lookups | contactable on open |
Questions
Frequently asked questions
How do you find off-market properties before anyone else?
Watch the public records that precede a sale. County recorder filings carry ownership and financial distress signals, liens, notices, transfers, that show up weeks before a property would ever be listed. A system that ingests those records nightly and scores them puts you in front of the owner while the window is still open.
What signals actually predict an off-market multifamily sale?
Ownership signals (long hold periods, out-of-area owners, recent transfers within a family or entity) and financial signals (liens, defaults, maturing debt indicators) are the recurring ones. No single signal is decisive; the scoring engine weighs 14 of them together and ranks properties by combined intent.
Why not just buy a list from a data vendor?
Vendor lists are the same lists your competitors buy, and they age from the day they are exported. This system reads the primary source directly and refreshes overnight, so the pipeline reflects what was recorded yesterday, ranked against the investor’s own buy box rather than a generic filter.
How much coverage does a system like this need to be useful?
Less than you would think. This engine went live with three counties and produced 1,800+ qualified leads a month, with the architecture built to scale to nine. Depth of signal per county matters more than raw county count.
Related
Keep reading
Work with us
Want a system like this?
We start with a paid AI audit: we map your workflows, identify where a system like this pays back, and scope the build before any larger commitment. If training your own team fits better than software, that is one of the three consulting offers.