Manufactured housing, Upper Midwest
An off-market sourcing engine for manufactured housing
We built Meadowlark an off-market sourcing engine for manufactured-housing parks across six Upper Midwest states, classifying every county in the footprint by how its records can lawfully be reached before a single record was pulled.
The problem
Finding park owners was a six-step manual process, and the step it turned on did not work. Owner phone numbers came back five to ten per name, and roughly half of them were wrong. An acquisitions team was spending its week qualifying a contact list rather than a pipeline.
What we built
A county access map before any scraping
Every county in the footprint was scored on how its records can lawfully be reached: free statewide bulk or open API, free per-county portal, paywalled or automation-prohibited, and no usable online access. Not every county resolved to a single tier. Forty-two per cent of the footprint is confirmed reachable for free but not yet split between statewide and per-county access, and it is carried as its own state rather than rounded up into the better one. That is the map working, not the map failing: an unfinished classification published as unfinished is the difference between a coverage claim you can act on and one you find out about later.
A motivated-seller score, explained line by line
A deterministic score across three tiers of ownership and distress signals, weighted for the rural Upper Midwest: long tenancy, individual and personal-trust ownership, tax delinquency, absentee and out-of-state ownership, mortgage age, recorded liens and lis pendens, and single-park operators. Every score is reconstructable from the record it was computed on.
Public-record owner resolution, with the chain attached
The title-holding entity traced to a named person through filings that are already public, and the chain handed over with the name so an analyst can verify it rather than trust it. Where a chain stops, the row says where it stopped and stays in the pipeline marked unresolved.
Built to run on their infrastructure
The pipeline, the records and the scoring are built to live in the client's own environment, under their governance. Independence from us is the design, and it is stated here as the design rather than as a finished handover.
197
counties in the scoped footprint, across six states
County feasibility matrix, one row per county. Our delivery record.
205
sources cited building the county footprint
Cited sources across the six research notes of the delivered audit.
14
signals in the delivered motivated-seller score
Signals across three tiers in the delivered scoring model.