Which property and owner signals are most useful for identifying potentially motivated sellers?

A signal earns its place by naming a dated event in a specific file. A notice of default recorded on a stated day, a probate case in a numbered court file and a tax balance printed as of a delinquent date each describe something that happened to one identifiable person.[4][6] A vacancy flag or a propensity score reports what a model believes about a house from a distance.

PropStream lists skip tracing at 12 cents per contact and direct mail from 57 cents per postcard.[1] A 20,000-record segment built on a filter that separates nobody costs $2,400 to trace and another $11,400 to mail before one owner answers.

Three tiers, ranked by where the evidence came from

Tier one covers events somebody recorded or filed, including foreclosure notices, tax delinquency, probate or divorce cases, recorded liens and expired listings.[3][4][6][7][8] Tier two covers conditions a model infers from other data, including vacancy indicators, estimated equity and 0 to 100 propensity scores.[10][12] Tier three is intent an owner stated out loud, and it is the only tier where the evidence came from the person who owns the house.

Tier one carries the most weight in daily use because it is dated and reproducible from a public file. Tier two needs a measured result on a held-back segment before it earns budget.[13]

Freshness decides sequencing more often than the label on the checkbox does. Two rows carrying the same signal name can be a week apart or a quarter apart, because a recorder indexes daily while the federal vacancy file aggregates quarterly.[9][10] PropertyRadar describes its distress and life event data as sourced from county courts and recorders,[2] so one checkbox in that catalog can rest on a court docket or on a model, with nothing in the export marking which is which.

Foreclosure is the densest tier-one signal, because the filing is the event. PropertyRadar covers judicial filings including lis pendens and sheriff sales alongside non-judicial notices of default and trustee sale notices, advertises preforeclosure listings, auction records and bank-owned records in every state updated daily, and sells optional trustee sale tracking across 270 counties that rechecks each scheduled property every 15 minutes on auction day.[3] A 6-week-old notice with no sale date is a different conversation from an auction scheduled Thursday.

Tax delinquency carries two dates, one from the collector's records and one from the day the list was printed. Mecklenburg County's monthly delinquent list states the tax record date it reflects, warns that a payment in full or an appeal appears only in the next list, and names owners as of the January 6, 2026 delinquent date.[4] Severity stays invisible on a checkbox, so the years-delinquent count separates the two ends of that filter, and most jurisdictions initiate a tax sale once taxes are 5 years delinquent.[5]

Probate is a court process before it is a data field, and California's self-help guide describes it as the legal process required to transfer or inherit property after the owner has died.[6] A personal representative is appointed to collect the property and settle the bills, and whether an estate reaches court depends on the amount and type of property involved, which is why a probate filter returns a fraction of the deaths in a county.[6]

An expired listing runs on a trade rule instead of a statute. The National Association of Realtors requires a broker to submit a listing to the MLS within one business day of marketing a property publicly,[8] while a divorce filing sits in family court, where a judge decides money and property questions along with children.[7] Both filings name a dated case while leaving the outcome for the house to be decided later.

Every tier-one signal inherits its county's indexing lag, and Harris County's real property search tells users a processed document may take 1 to 2 business days to appear in a search inquiry.[9] A list you pull Monday morning is missing Friday's recordings.

A vacancy flag reports 90 days of uncollected mail rather than an empty house. The Postal Service counts an urban address as vacant once delivery staff have identified it as not collecting mail for 90 days or longer, and those counts reach HUD as a quarterly aggregate.[10] PropertyRadar's criteria glossary defines Site Vacant by that postal determination and says the data is made available to suppress mailing to undeliverable addresses.[5]

Building a target list on a suppression field inverts what the field was made for. PropertyRadar's vacancy guide notes that flagged properties can still receive mail while nobody picks it up, names short-term rentals such as Airbnbs as a known false positive, and tells anyone targeting a vacation destination to expect results that need confirming.[11]

Estimated equity is arithmetic run on one recorded number and one modeled number. The criteria glossary works the example with a first mortgage of $50,000 and a second of $30,000 against an estimated value of $100,000, giving a combined loan to value of 80% on the same parcel that shows 20% estimated equity.[5] The recorded half is exact and the modeled half is not, which is why the glossary labels the result estimated.[5]

Propensity scores are the tier-two entry most often mistaken for intent.[12] PropertyRadar's Likely Transactions puts a 0 to 100 score on every property across likely to sell, purchase, refinance and HELOC signals.[12] The bands run from Very Low to Very High, estimating who is most likely to transact in the next six months, with every property rescored monthly.[12] Its chief executive is quoted on that announcement calling the feature "not a crystal ball",[12] and the feature page adds that likely scores are directional, model-based estimates for marketing use only.[2]

A score earns budget by producing a measured result on a segment held back from outreach. The metric is precision, defined in Google's machine learning course as the proportion of a model's positive classifications that are actually positive.[13] Here that becomes the share of top-band properties that transacted inside the score's window, which takes 6 months to observe.

Verified intent starts when an owner answers a mailer, replies to a text or gives a caller a timeline. It is the scarcest tier because outreach creates it instead of filtering finding it, and it lives in whatever the caller typed into a note.

An owner who says do not call again has stated intent as clearly as one who asks for an offer, and that version carries legal weight. The FTC's Telemarketing Sales Rule makes it a violation to call any consumer who has asked not to be called again, and puts responsibility for keeping a current entity-specific Do Not Call list on the seller.[14] The FCC rule requires that request to be honored within 10 business days and kept honored for 5 years.[15]

Recording intent the moment it appears is what stops a list from dialing the same owner every quarter. A suppression flag written into the record on the day of the call carries the 5-year obligation forward into the next campaign.[15]

Two signals from two different offices carry more weight than two signals pulled off the same document. Estimated equity and length of ownership both derive from the same recorded deed and the same recorded loans,[5] so a list filtered on both has applied one fact twice and narrowed less than the count suggests. A probate case and a vacancy flag come from a court clerk and a letter carrier, and neither observer knows the other exists.[6][10]

A BiggerPockets poster working vacant properties described stacking vacancy, unpaid taxes or defaulted loans, and a deceased owner.[16] That draws on records from 3 separate offices, though no forum thread can establish that the combination converts better than another.

Narrowing before you buy contact data is where the tier ranking turns into money. PropStream charges 12 cents per contact for skip tracing,[1] so cutting a 12,000-record segment to 900 by requiring 2 independently sourced events is the difference between a $1,440 trace bill and a $108 one.

PropertyRadar's criteria glossary lists 250+ search criteria available for building a list.[5] Filter menus flatten that catalog into checkboxes, and the same catalog carries distress and life event fields described as sourced from county courts and recorders.[2] A court filing and a modeled estimate arrive in that menu with no marker separating them.

Catching a dated event on the day it lands is what monitoring buys, and PropStream's Lead Automator watches up to 50,000 properties on its bundled plan.[1] That re-evaluates a saved set instead of waiting for somebody to re-run an export. Monitoring pays on tier-one signals specifically, since those are the rows where a date changes.

County tax collector delinquent lists and recorder indexes cost nothing beyond the time to work them.[4][9] What a platform adds is assembly across thousands of counties, which changes the cost of the work rather than what the filing proves.

Absentee ownership misleads more often than its reputation suggests, because the filter compares two address fields.[5] The criteria glossary notes that some areas have no mail delivery, so an owner's mailing address can differ even where the owner occupies the property.[5] A property manager, a CPA's office or an adult child's house produces the same mismatch as a landlord 3 states away.

Estimated equity errs in the opposite direction, since a modeled value that ignores condition can show comfortable equity on a parcel needing a new roof and a foundation.[5] A recorded junior lien or tax lien never enters a combined loan to value calculation at all.[5] A 20% equity figure and a 20% equity fact are therefore separate things.

A published percentage can borrow the authority of a study nobody names, and SmartZip's homepage is an example.[17] It markets a listing-prediction product on the promise of predicting listings and supports the pitch with a claim that 70% of homeowners choose the first agent they meet.[17] None of the signals described here carries a published predictive percentage, so a number in that shape stays a marketing claim until somebody names its denominator.

Harris County states that its online database may not reflect the complete or unaltered contents of the official records.[9] The offices holding these records say they carry errors, and a misspelled owner name breaks the join between a court filing and a property row, which produces rows that look clean and describe the wrong house.

A filter is not motivation. Checking a box for high equity, long ownership or absentee status selects properties sharing an attribute that says nothing about wanting to sell.[5] The count that comes back measures how many parcels match a rule, not how many owners will answer.

A score is not intent either. PropertyRadar describes customers filtering by score, sorting high to low, and layering scores on top of criteria such as equity and length of ownership,[12] which places a band beside a dated filing rather than above one.

A vacancy flag is not an abandoned house, since the field exists to suppress undeliverable mail and a short-term rental produces the same flag as a boarded-up property.[5][11] A tax delinquency checkbox is not a pending tax sale, because most jurisdictions wait until taxes are 5 years delinquent before initiating one.[5] Both filters return real records that leave you to supply the urgency the record itself does not contain.