Have suggestions for products that you’d like to see reviewed by our real estate technology expert? Email Craig Rowe.
Using blends of public and private data to predict buyer or seller behavior is a common feature within customer relationship management or digital marketing tools.
The majority of these scoring systems are anecdotal, however, using minimally gathered instances of home searches or social media likes and activity. They also tend to be secondary features to a product’s greater purpose.
Revaluate, after launching as a neighborhood ranking tool, has since recentered its business model on analyzing online consumer behavior, which churns hundreds of data points through its machine learning algorithms.
To verify its efforts, the Golden, CO company announced last week that it commissioned Altos Research to perform an independent study of its scoring accuracy.
The study’s results showed the company’s “Likely-To-Move” label is 19.5 percent accurate, meaning 1 out of 5 leads it scores will sell their home within six months.
Subjects of the study were pulled directly from an MLS, and renters were removed. According to a press release exclusive to Inman, Altos used a statistically significant sample size (5 percent) of all the likely mover leads generated by Revaluate in 2017 and verified them against closed sales.
In the press release, Revaluate’s CEO Chris Drayer claims customers report a higher percentage of certainty.
“Our customers currently self-report that Revaluate’s prediction success rate is 36.4 percent, which would account for the renters and homebuyers that Altos Research could not track in its study since they could not be independently verified,” Drayer said.
Inman first reported on the 36.4 percent rate in October 2018.
The most recent test used subjects from throughout the United States, excluding New York City and San Francisco, where MLS data isn’t as widespread. It also left out first-time homebuyers (who don’t have a house to sell) and second or vacation homebuyers.
“This is by far the best lead success rate we’ve seen in the real estate industry,” said Mike Simonson, CEO of Altos Research in the release. “We expect the rate to improve over time, given the machine learning Revaluate’s algorithms employ.”
One part of Revaluate’s data model focuses on what it calls the “D’s:”
- Diamonds: marriage and engagements
- Diapers: birth announcements
- Daily grind: job news
- Diplomas: children graduating from college
- Divorce: separation of families
- Death: in-residence family member dying
- Discretionary income/Downsizing: retirement or
purchase of second home
Online data collection and processing is typically juxtaposed with the accuracy of cold calling, considered only 1 percent successful.
Altos states that Zillow, Trulia and realtor.com leads have an approximate 10 percent success rate, according to the press release.
That result came from the review 5 million homes sold and 50 million leads generated.
Have a technology product you would like to discuss? Email Craig Rowe