AI Case Study

"The couple was motivated by the 2 percent commission they would pay the artificial intelligence-driven service run by REX Real Estate Exchange instead of the typical 5 to 6 percent paid to a human broker who would find buyers for them.

The bidders for the Hougardys’ home weren’t found by an agent adding the home to a multiple listing service, which feeds that information to Zillow, Trulia and Realtor.com.

Instead, computers were crunching over a hundred thousand data points in order to identify thousands of likely buyers and target them with ads on websites and social media.

“We can tell everybody who lives within five miles of your house who has the basic means and motivation to want to buy your home,” said Jack Ryan, co-founder and CEO of REX Real Estate Exchange, based in Woodland Hills, California.

He says his computers find potential home buyers based on who is clicking on REX ads, as well as recent purchase decisions people have made coupled with their history of home ownership.

That’s resulted in 231 closed transactions since REX’s launch in 2016, according to the company. REX represents sellers as well as buyers.

To sell a home, REX computers disseminate an initial batch of ads based on a hypothesis about where and whom the likely buyers are.

Of the people who clicked on the ad, “the computer in real time, like right now, can figure out what those 500 people have in common,” Ryan said. “Let’s go find more people who look just like those 500 people who clicked on the ad.”

If someone clicks on an ad, the computer will tailor future ads for that individual based on what their behavior on the website demonstrated about their interests.

REX also identifies potential home buyers and sellers by gathering data from retailers and businesses that show changes in people’s purchasing behavior. It’s common practice for businesses to sell data about consumers’ purchases to third parties and advertisers, privacy experts say.

“The fact that you bought a hammer itself isn’t very informative to us,” Ryan said. “The changes in the rate at which you’re buying home supplies is very important to us.”

That’s coupled with data about the individual’s homeownership history – whether he or she actually owns a home and has positive equity in the home in order to make a down payment on a new one. Likewise if someone just purchased a big screen TV, it’s likely they aren’t about to sell their home or move into a new one.

REX also tailors ads by determining which features about a home or its surrounding community can increase its probability of selling faster or at a higher price point, Ryan said.

Computers analyze data about comparable home sales in the area. Then REX data scientists input the facts about the home for sale and introduce new variables to determine how they affect the probability the home will sell for a certain price in a certain number of days.

“It says to us, hey don’t know why but apparently the probability of success goes up by 3 or 4 percent if it’s within 2 miles of a Starbucks or something,” Ryan said. “There are dozens of pieces of data, each of which changes the probability by one or two percent.”