Covering Disruptive Technology Powering Business in The Digital Age

Home > Archives > News > Marketing Tool Uses Big Data to Predict Home Buyers and Borrowers
Marketing Tool Uses Big Data to Predict Home Buyers and Borrowers
July 26, 2016 News


Just as predictive analytics can entice Amazon customers into additional purchases and Microsoft, armed with Big Data, can predict pancreatic cancer diagnoses from web searches, mortgage lenders can use similar tools to zero in on likely future borrowers.

One such tool, SmartZip, is used by about 3,500 real estate professionals. The predictive marketing software for real estate is at least two times better at picking homeowners who will need a real estate agent than choosing potential clients at random, the Pleasanton, Calif., company claims.

“We’ve spent the past seven years helping agents and brokers identify, rank and convert the homeowners (who are) most likely to sell in their neighborhood areas,” said Avi Gupta, SmartZip president and chief executive officer. “We know that real estate professionals can benefit from patterns and signals that go unnoticed to the naked eye but are easily identified by a smartly programmed algorithm.”

For example, if an agent’s farming area — the market area in which an agent tends to specialize — has 2,000 houses, the software can select the 400 owners who are most likely to sell over the next 12 months, the company claims. Of those owners, 40 to 48 — or 10% to 12% — will actually sell.

“You’ve got a 1-in-20 chance in finding a potential seller client if you randomly knock on doors,” said Peter Grace, a SmartZip vice president. “But you’ve got a 1-in-10 or -9 chance if you knock on the doors we tell you to knock on.”

For lenders, SmartZip works a bit differently. Besides analyzing a particular area to search for owners likely to sell, often in partnership with real estate agents, it also uses predictive analytics to provide a “pre-mover score” for every contact in the lender’s database.

The more likely someone is to sell or move within the next six months, the higher their score. And based on the score, lenders can segment the top prospects and allocate their marketing and sales efforts accordingly.

Sean Stanfield of HomePort Financial Services in San Diego partners with 40 agents in his market area. “Only 10% use SmartZip, but I wish more did,” he said.

The SmartZip program, SmartTargeting, runs through about 2,000 variables to make its predictions, including the year the houses were built, the size of the property and the farm area’s migration pattern, including how many houses sold over the last 12 months. It also zeroes in on each owner’s number of children and their ages, each owner’s estimated net worth, how long they’ve lived in the home and details of their current loans, including their loan-to-value ratios. The variables are different for every market area, which can be as large as a single ZIP code or as small as a Census Bureau block.

In a recent test of the program’s accuracy, SmartZip went back 12 months and focused on a 2,400-home section of San Jose, Calif. Had the program been working for an agent, it would have told the agent to focus marketing on 493 particular houses. Of those, 9.2% actually sold versus 4.6% for the entire market area, or about 45 people.

SmartZip’s selling point: Those 45 sellers need someone to list their homes for sale and perhaps an agent to work with in finding a new house. And the sellers who are looking for a new house will need a new mortgage.

In addition to finding better leads, SmartTargeting can help users identify the farm area with highest turnover rate, or the area with the highest commission potential.

In a farm area with 1,500 homes, no agent is going to knock on the doors of 300 homes — the top 20% — let alone all 1,500 homes, Stanfield said. But if the agent can call on 2% of those top 300 owners — six people — it’s manageable.

Once the algorithm identifies the best prospects, the program can help reach them with a multi-channel marketing program that produces emails and direct mail pieces, including digitally hand-written letters and post cards. The program also produces a branding piece that pops up every time a prospective client goes to Facebook and other major websites.

The idea is to stop wasting marketing on people who aren’t interested, and instead launch targeted campaigns aimed at prospects with the highest moving scores.

This article was originally published on and can be viewed in full