Raleigh, NC / Sunnyvale, CA, May 19, 2022. One of the best stories to emerge from RESO’s recent 2022 Spring Conference was how MLSListings, based in the heart of Silicon Valley, stepped to the forefront in implementing the RESO Universal Property Identifier (UPI) standard into its multiple listing service (MLS) system.
Joel Allison and Olga Ermolin of MLSListings.
According to Olga Ermolin, Director of Engineering at MLSListings, “We were determined to do something beyond a simple proof of concept, intentionally implementing the UPI into full production to solve real issues right now, such as delivering thorough property history to our subscribers.”
The issues in question, as described during a presentation to nearly 300 conference attendees, included the expulsion of duplicate listings, a desire for more accurate market statistics and a complete view of property history independent of the MLS in which the sale took place. | VIEW PRESENTATION SLIDES | WATCH PRESENTATION VIDEO
“While we have been diligently working toward modernizing real estate data transport with the RESO Web API and creating transparency with RESO Analytics, we were absolutely delighted to learn that one of our MLS members has implemented an entirely different but no less important standard with the UPI,” said RESO CEO Sam DeBord.
The effort created clean, complete, reusable property data that, in turn, allowed MLSListings to release a true data share with seven neighboring MLSs. Combined with data shares in non-bordering parts of California, more than 100,000 duplicate and triplicate listings have been purged from their system.
MLSListings determined that the only way to severely reduce duplication was through a unique ID. One option, the Assessor Parcel Number (APN), a unique ID assigned to land parcels by a county tax assessor, proved to be too unreliable for the task on its own. They could have created an internal ID, but they decided to use the RESO UPI standard to help pave the way for wider integrations.
In all, the effort covered 8.9 million listings and the assignment of 8.4 million UPI numbers for a coverage of 94 percent. Perfect UPI matches accounted for 83 percent of listings, while partial matches accounted for 11 percent and internally crafted “synthetic UPIs” were created for 6 percent of listings. APN accuracy amounted to 87 percent and address accuracy landed at 89 percent. An accuracy score was developed by MLSListings to complete the assignment.
Summing up the whole of the project for future implementers, Joel Allison, Senior Software Engineer at MLSListings, stated, “Exact matching is a luxury, not a necessity. Failure is an option.”
Future plans call for the inclusion of latitude/longitude information, integration with other data sources, the attachment of APNs, visualizations like market-driven heat maps and enrichment of data related to the sale of the property (e.g., appraisals, liens, loans, offers and permits).
About MLSListings
MLSListings provides real estate professionals and consumers with accurate data that is updated every five minutes. Facilitating more than $70 billion in annual real estate activity, the MLSListings platform is the intersection of comprehensive real estate data and the transaction for the northern California marketplace. | mlslistings.com
