Vacant House Data Feed: How Wholesalers Find Abandoned Properties
Many vacancy flags trace back to a mail carrier's judgment, which is useful and limited, so here is how to tell a good feed from a noisy one.
Tennessee, Goliath Teammate
Key Takeaways
- A vacant house data feed is a recurring supply of addresses flagged as probably empty, delivered as a file, an API or a saved list in software, and many flags trace back to U.S. Postal Service delivery data.
- HUD counts an urban address as vacant when USPS delivery staff identify it as not collecting mail for 90 days or longer; vacation areas and high-growth areas inflate vacant and no-stat counts.
- USPS DSF2, the licensed product that contains vacant address information, carries an annual license fee of $230,800, so investors typically reach the data through vendors such as PropStream, DealMachine, PropertyRadar or Smarty.
- A vacant flag alone does not prove a house is empty: confirm by driving a sample of 25 flagged houses in one ZIP code, and stack a second signal such as tax delinquency or probate.
- Goliath is not sold as a USPS vacancy product; plans start at $1,499/mo and include 5,000 to 25,000 skip traces a month.
Goliath Data publishes this guide, and Goliath is one of the services compared. Prices and features were checked on each company's own site on November 5, 2026, and can change.
A vacant house data feed is a recurring supply of addresses flagged as probably empty, delivered as a file, an API or a saved list in software. Many vacancy flags trace back to the U.S. Postal Service, often with city and county records layered on top. A good vacant house data feed tells you where to look, not what is true. You still confirm each house before you spend a dollar on it.
The useful question is not "who has the biggest feed?" It is "what is each flag based on, and how old is it?"
Where does a vacant house data feed get its flags?
USPS delivery status. USPS offers this information through licensed products, and the price tag explains a lot. The DSF2 product, which contains vacant, residential, business and seasonal address information, carries an annual license fee of $230,800, and applicants must be in the address verification and list processing business. So investors typically reach this data through vendors.
City records. Some local governments publish their own indicators. Philadelphia's Vacant Property Indicators dataset identifies properties likely to be vacant from city administrative data, and it is offered as CSV, GeoJSON and an API, which is about as close to a free feed as you will find. Chicago publishes a dataset of vacant and abandoned building violations. Philadelphia's own terms say it makes no representation about the accuracy of any specific information, so verify before you act. Search your own city's open data portal first, since it costs nothing to look.
Signals that often travel with vacancy. Tax delinquency, absentee ownership, probate and pre-foreclosure do not prove a house is empty. They do explain why it might be, and they suggest who might want out.
How do you read a USPS-based vacant flag correctly?
HUD's description of USPS data sets the expectations. An urban address counts as vacant when delivery staff identify it as not collecting mail for 90 days or longer. Rural route addresses with the same pattern can land in a different category called no-stat, along with homes under construction.
That leads to three practical traps:
- Vacation areas. HUD notes that vacation and resort areas have very high vacant address rates. A beach condo with no mail in February may just be a beach condo.
- Growth areas. HUD says high-growth areas show high no-stat rates because of new construction. Those houses are not abandoned, they are unfinished.
- Stale statuses. The address API company Smarty warns in its documentation that USPS is often several months behind in removing addresses from the no-stat list, and it advises relying on that field only if you know its limitations.
USPS also records nothing about why an address is vacant. To USPS, it either needs mail service or it does not. Motivation is your job, and our piece on reading the motivation behind vacant properties gets into that.
Which vacant house data feed options are there?
Goliath is listed first on neutral criteria (signals beyond vacancy, source kept on each record, outreach included). It is not sold as a USPS vacancy product, so if a pure vacancy flag is the job, look at the others.
Service | What the vacancy data looks like | How it arrives | Price, checked November 5, 2026 |
|---|---|---|---|
Related signals with source shown, not a USPS vacancy list | Inside its CRM | From $1,499/mo | |
A "Vacant" list: properties marked vacant by USPS | Saved lists and exports | ||
Vacancy status filter: vacant, likely vacant, occupied | Search with 800+ filters | ||
A "Vacant" list | Lists, exports, API access | From $119/mo | |
`dpv_vacant` yes or no per address you submit | API | 42-day free trial, then from $552/yr for 12K lookups |
The Smarty approach is different in kind. You bring the addresses and it returns the status, so it suits a developer who already has a target list and wants to screen it. Regrid and ATTOM sell property data by API and bulk delivery too. Ask their sales teams which vacancy fields you would get before assuming.
What should you ask any vacant house data feed vendor?
- What creates the flag? USPS status, a model, or a person? "Likely vacant" needs a definition.
- How often does it refresh, by source and in my county? A quarterly update and a weekly one are different products.
- Do I get the date the status changed? Fresh vacancies and ones sitting for two years call for different scripts.
- Can you separate seasonal and no-stat addresses? If not, expect resort and construction noise.
- How does it arrive? CSV is fine at low volume. Automation needs an API or a scheduled delivery.
- What may I do with the data? Read the license, especially if you plan to resell or share it with partners.
How do you test a vacant house data feed in your market?
Vendors will not agree on accuracy, and we have not seen an independent study we would quote. So measure it yourself. Take a sample of 25 flagged houses in one ZIP code and drive past them. Note overgrown yards, boarded windows, no furniture and full mailboxes, using these visual signs. Stay on public property while you do it.
Count how many look empty, then try a second vendor on the same ZIP. Your own hit rate beats any number on a sales page, and running it costs a tank of gas.
How do you turn a vacant house data feed into seller conversations?
A flag alone does not get you a deal. Build a small routine around it.
Remove what you cannot use. Filter out your own past contacts, and any property type outside your buy box.
Stack a second signal. A vacant flag with delinquent taxes or a probate filing often says more than a vacant flag alone. Check both against county records before you make an offer.
Mail the owner, not the house. Use the owner's mailing address. Sending a letter to an empty house rarely reaches anyone.
Skip-trace the owner. Skip-trace tools report similar match rates, around 99%, so the difference is accuracy. Goliath publishes a 99% match rate and 85% accuracy with no charge for misses, and every vendor should show you how it measures this.
Track outcomes. A CRM lets you record who answered and stop repeat calls. Inside Goliath, David can handle calls and configured follow-up, then pass the seller's history to your team. Plans include 5,000 to 25,000 skip traces a month, so it fits a team already doing volume.
Before dialing, read the FTC's Do Not Call guidance. Whether an offer to buy a house counts as telemarketing is a question for your lawyer, so treat calls and texts as covered. This is not legal advice.
Should you pay for a vacant house data feed at all?
Check your city's open data first. If you work a small area and a saved filter in a $99 tool shows enough houses, a dedicated feed may be overkill. Consider an API once you have a developer and enough volume that manual exports waste real hours. Driving your target neighborhoods remains the cheap way to confirm all of it.