For all the talk about how Zillow’s iBuying fiasco was a triumph of humans over machines, it’s worth remembering one simple fact: Someone had to tell the bots which data to use.
Agents and rival homebuyers are crowing about how Zillow’s apparent overreliance on a faulty algorithm led it to pay thousands over market prices. They say it shows the importance of on-the-ground expertise in a business that perhaps more than any other depends on human-to-human interactions.
“It reinforces the need for a local area expert when you’re buying or selling a home,” said Babbi Gabel, a top agent in the busy iBuying nexus of Phoenix. “Real estate is about relationships.”
Yet the answer may be more nuanced. After all, competitors have long used computer programming to identify, buy and sell homes at a profit. Rather than some sort of post-apocalyptic tale of how ragtag rebels faced down Skynet, it’s more the story of how hubris overtook good judgement.
“Investors should ask questions about execution,” said Tom White, an analyst at D.A. Davidson & Co. who rates Zillow a buy. “They clearly screwed up.”
Zillow Offers did use pricing specialists and occasionally agents to help resolve discrepancies in property values and to find valid comparative prices, a person familiar with the company said, speaking on condition of anonymity. Yet Zillow relied mostly on its much-touted, and long-derided Zestimate and stripped out most human interaction.
Rivals such as RedfinNOW and HomeVestors, by contrast, built in human checkpoints to review automated prices and underlying assumptions. In some cases, they inspect a home they’re considering buying. Offerpad’s CEO Brian Bair said in a statement that it relies on “local market expertise,” without elaborating. Opendoor says on its web site that it hires real estate professionals to man pricing teams.
RedfinNOW uses extensive renovations to deliver upside to ensure it can sell a home for more than it paid. Its specialists also manually review Redfin’s own comparative analysis for every offer it makes. Redfin’s licensed home inspectors visit every home and take 3D scans of the property.
“Those are things that we think humans are really good at,” said Quinn Hawkins, who heads RedfinNOW.
Orchard, a startup that provides loans based on a previous home’s value and will manage its sale, uses an automated valuation system that relies on data from public records, a structured interview with the homeowner, a virtual tour and photos. Yet that’s only 70 percent of the process. The rest of the time, or when there aren’t enough comparable properties, humans step in.
“The key is the system needs to be self-aware to know when it’s confident and when it’s not,” said Orchard CEO Court Cunningham.
Then there’s Zillow’s own obsession with growth.
“We made a decision that we needed to increase our acquisition pace” even as the pricing model failed to keep up, Allen Parker, Zillow’s CFO, said on a conference call Tuesday with analysts.
Zillow’s stumble probably almost certainly isn’t the end of the road for iBuying. Younger people, more at ease with clickable purchases, may seek the path of least resistance when it comes to buying and selling homes.
Even so, Zillow’s sudden exit spooked investors. Its shares slid 32 percent this week bringing the loss since a peak in February to 67 percent.
Trulia co-founder Sean Black, who sold his site to Zillow for $2.5 billion in 2015, said he’s rooting for the company. The stock’s dramatic slump this week is a buying opportunity provided “you have the stomach for the volatility,” said Black, who holds Zillow shares.
Even so, Black said home pricing needs a human to conduct a qualitative review.
“Homes are not a commodity,” said Black, who has since founded home finance startup Knock. “You need a human element to it.”
The post Humans vs machines: The fallout from Zillow’s iBuying fiasco appeared first on The Real Deal Los Angeles.
Powered by WPeMatico