"Firms know when you're going to move before you’ve decided you're going to move - the data exists"
Predictive modeling is the future for the real estate industry, and mortgage professionals can either embrace the technology or lose business, Brad Sivert (pictured), head of marketing and proptech at Tavant has told MPA.
The comments come after the Silicon Valley-based firm released a 12-page whitepaper on consumer behavior using machine learning (ML) to predict the likelihood of a homeowner selling their home in the next 12 months.
As one of the report’s authors, Sivert explained the reasoning behind it, saying there was a need to empower real estate agents and buyers with the necessary information to know when a home was going to sell.
He said: “There’s enough data in this world for the most part. Amazon probably knows you’re getting married before you actually propose to someone. They (also) know when you’re going to move before you’ve decided you’re going to move. The data exists.”
Tavant’s team discovered that an effective prediction model would help agents engage with the right homeowners at the right time.
Read more: Don’t fear the tech, head tells feet-dragging brokers
“In a market like ours, homes are going quick. We need to empower both the real estate agent and the buyers with information to know when a home is going to sell,” he added.
Tavant’s predictive ML model included data based on three categories: property details and homeownership data – including behavioral, family and financial information - as well as miscellaneous details involving price comparisons of homes in geographical proximity.
The ML model also gauged the reasons for selling a property, including changes in household size or increases in the neighborhood’s population, or simply a desire to upgrade or downsize a home.
Five US counties - Marion, Brevard, Ventura, Arapaho and Clark – were chosen for the modeling exercise due to having “the cleanest data” and the highest likelihood of accuracy.
Sifting through the data, Tavant checked a total of 113 attributes - of those, about 10 were deemed the most important, as they determined the likelihood that someone was going to list and sell their home.
Sivert added: “Age is a huge factor; where your children are, the valuation of your home versus what you bought it for - not necessarily how much equity you earned but what the perceived value is.”
Sivert said it was now possible to predict when a home was going to sell with 50% accuracy, adding that predictive modeling could look as far as 180 days ahead.
He said: “There’s a lot of value in that for the real estate agent, the home seller, the home buyer, and it also tells you the difference if you’re looking to list your home in August versus listing it in October.”
Sivert argued that predictive modeling was here to stay, despite the recent collapse of a similar model as used by online real estate firm, Zillow.
Earlier this month, the company announced that it would be closing its ‘Offers’ homebuying side of the business after reporting a $422 million loss during Q3.
Read more: Zillow’s fix-and-flip ‘Offers’ business to close
Sivert stressed that Zillow’s failure had been due more to a faulty business model rather than to ML tech, saying that people ended up buying homes for less than Zillow had acquired them for, although he also posited that the firm’s tech experts had been unable to hone their algorithms.
Sivert however recognized that predictive modeling had not taken an equally important variable into account – the emotional reasons for buying a home.
“That’s the part that no-one’s really been able to crack because that’s emotion, and data can’t really define emotion,” he said.
That “unquantifiable aspect” meant that a homebuyer determined to acquire a property for no apparent logical reason would be prepared to pay 2% above the asking price.
He said: “American mentality is to maximize your debt because you want that dream home. These emotional aspects are the hardest part to quantify when data comes in, because they’re different for all people.”
Asked if mortgage professionals should feel threatened by predictive modeling, he replied: “Real estate agents can either feel threatened by this amount of data…or they can use it; incorporate it into their business and do twice as much business.
“Either way, the data will continue to be used. It just depends on what side the real estate agent wants to be on. Either the fear side or the embrace side.”
