Listing data is the lifeblood of the real estate industry. Innumerable websites and apps run on it, benefitting agents, brokers, consumers and the companies that thrive on providing services to them.
But what goes on behind the scenes to make ubiquitous data a reality?
The days of MLS books are long gone. Now, multiple listing services transfer data to vendors chosen by brokers and agents through two main data standards: the Real Estate Transaction Standard (RETS) or an application programming interface (API). They each have their benefits and challenges, but the latter is the future, according to panelists on Tuesday at Inman Connect San Francisco’s Hacker Connect.
If a real estate tech company has been around for more than a decade, chances are they’re using RETS. But it’s “pretty old technology,” said Seth Siegler of Hoverboard Labs during a session called “Dealing with the MLS: A Developer’s Perspective.”
Siegler, who used to be a real estate broker and was raised by an agent mom, lamented the different configurations of the more than 600 MLS systems nationwide, requiring software developers to take on complex data “normalization” for each MLS they integrate on behalf of their broker clients.
“Expect the unexpected,” Siegler advised developer attendees. “You’re going to see things that you wouldn’t even believe. These things were created a long time ago. Some of them have things broken up in different tables, single family and multifamily in different tables …”
His “dream” would be for all MLSs to ditch RETS and adopt the standardized web API from the Real Estate Standards Organization (RESO). Right now, it would be a lot more work to integrate multiple MLSs through the API and others using RETS, he said.
Firms taking in RETS feeds must typically import data from the MLS nonstop, which Siegler said “makes absolutely no sense” given that MLS data is always changing. With the RESO Web API, tech firms can pull in only the data they need as a query is made in real time.
When both MLSs and real estate tech firms adopt RESO standards, the latter don’t have to spend as much time and money coding their products to meet the data peculiarities of each individual MLS — potentially resulting in more and better tools available to more agents and brokers, faster.
APIs have their own challenges
That’s not to say that APIs are foolproof. In a session called, “Building Smarter APIs: How Can We Standardize APIs for Real Estate,” panelist Haitham Al-Beik, founder of real estate startup letsbutterfly, noted, “There’s many redundancies with APIs themselves. Some companies try to add too many features.”
In addition, not everyone uses the same type of API, noted Aakriti Agarwal, product manager at consumer lending startup Blend. Some developers use SOAP APIs, some use REST APIs or others (RESO’s Web API uses the REST format).
Agarwal and Howard Hsieh of real estate computer vision firm Restb.ai urged developers to have documentation of their APIs that include examples of how their API can be used as well as failure scenarios that let users know how and when the API will “break.”
Moreover, a developer should write the documentation or at least look at it, Agarwal added.
Restb.ai tags and classifies specific features in property photos, identifying when a photo pictures a living room with hardwood floors, for example. Al-Beik pointed to Restb.ai’s API, which uses the RESO API standard, as an example of an API that does one thing and does it right.
“That allows me flexibility so that I can design my UX [user experience] however I want, knowing they can only do one simple thing really really well,” Al-Beik said.
Hsieh said the startup made its API “as simple as possible” with the intent to “layer in the complexity and robustness as we grow.”
Agarwal advised API developers to “Start simple. Don’t over-engineer and accept the fact that change is going to happen.”
Befuddled by APIs? Help is available
The National Association of Realtors required Realtor-affiliated MLSs to implement the RESO Web API standard on June 30, 2016, but many MLSs did not comply. Some real estate tech firms, such as CoreLogic and Bridge Interactive, provide products that can bring MLSs up to speed.
In a session called “Accessing the RESO Web API,” Jay Lee, a manager at Bridge, observed that once a product is built to the RESO Web API specification, “you can be confident it’s going to work” across the country, from San Francisco to New York.
Developers for tech firms and MLSs should be sure that the standardization platform they’re working with is compliant with the highest RESO standards, according to Lee. “I think there’s a benefit for you as developers in terms of future-proofing yourselves,” he said.
Developers should also ask whether the platform supports any MLS system, whether it has a dedicated support team, whether the platform offers additional data sets such as public records or the Zillow Zestimate and how much these cost, and whether the platform can handle everything from a live, real-time query to replication of the entire MLS database, according to Lee.
Kevin Regensberg, product owner at Bridge, demonstrated the company’s platform through its API Explorer tool, which lets developers “get a feel” for what its like to work with the Bridge API itself.
Developers can use the tool to play with their own data, but they can also use a free, standardized MLS data feed from the Austin Board of Realtors, which is what Regensberg demonstrated at ICSF.
Bridge’s platform pulls data from the MLS every 15 minutes. The company gives MLSs the choice of whether to allow full replication of their database, Regensberg said. There are some use cases, such as analytics or a big portal such as Zillow, where replication is “where it’s at” rather than live queries, he added.
“MLSs sometimes don’t allow you to replicate by default because they want to be good stewards of the data,” said RESO CEO Jeremy Crawford. “But most MLSs will allow [it] with a proper data license.”