Opendoor, the San Francisco-based real estate tech firm that uses artificial intelligence techniques to offer homeowners cash money for their homes, announced Tuesday it's now serving New York and New Jersey in select counties. This is the nine-year-old company's 46th market to date.

"This speaks to our continued ability to scale our operational platform and our customer experience, nationwide," said Ian Wong, Opendoor's co-founder and chief technology officer, in an interview withZDNetvia Google Meet. 

Recommends

The 7 best smartphones

From 5G to an amazing camera -- there's a phone here to meet your every need.

Read now

Opendoor said it will buy homes on Long Island in New York, and the lower Hudson Valley area of the state, including Nassau County, Orange County, Rockland County, Suffolk County, and Westchester County. 

In New Jersey, Opendoor will concentrate on the North and Central Jersey areas, including Bergen, Essex, Hudson, Hunterdon, Middlesex, Monmouth, Morris, Passaic, Somerset, Sussex, Union, and Warren counties.

The company will be purchasing single family homes only in a price range of$300,000 to$950,000.

"It's a natural progression for us to expand to more and more complex markets, and deeper markets," Wong said.

The company is not yet buying in Manhattan or the other New York City boroughs. "Our goal is to be there are some point," said Wong. The company is mindful, however, of the movement out of the city in recent years. 

"We see a lot of demand in these other zip codes," he said. "We want to make sure we can deliver a seamless experience to this set of customers first."

Opendoor

The New York and New Jersey markets have challenges that make them different from the 45 other markets Opendoor operates in, said Wong. 

"New York and New Jersey have very heterogeneous housing stock," he said. "These are places with a certain amount of history to them, there is a lot of variance in terms of the types of housing, what's in them; we've had to really leverage our capabilities in understanding home condition, normalizing the various data sets together."

The diversity in New York and New Jersey includes a "wider distribution of, say, home commission, layouts, number of bedrooms and bathrooms," said Wong. For any given home, "there is a sparsity of comparable" units.

Also:Opendoor CFO: 2021 was a 'breakout year,' and the 2022 housing market outlook is 'robust'

Wong explained that the variety of homes in the area "places an onerous burden to make sure that [Opendoor] can price homes well."

"This is the single largest transaction that most people do," he noted, "and if you don't give compelling offers to customers, they just won't accept. 

"A compelling offer is table stakes."  

That heterogeneity of New York and New Jersey is in contrast to the homogeneity of the Phoenix market, where Opendoor first focused on proving its business proposition years ago. "A lot of homes are newer, they have master plan communities, they are cookie cutter for many parts of Phoenix," Wong explained. That sameness minimized the risk for Opendoor's AI-driven pricing model. 

Adjusting to the complexity of New York and New Jersey meant "we have had to flex a lot of muscles we have developed over time on data ingestion, including from third-party vendors, and to normalize those data sets, including a lot of annotation." 

The heterogeneity of New York and New Jersey housing stock presents challenges for data science, says co-founder and CTO Ian Wong.

Opendoor

Opendoor is not just collecting data, Wong emphasized. "We are also spending a lot of energy, in-house, to provide first-party overlays, annotations, and augmentation of data sets to make sure we have a comprehensive look at the housing stock -- to make sure we can price them and give compelling service to our customers."

Also:Opendoor discusses the secret sauce: 'A deeper mechanism to the world'

Because of the sparsity in the New York and New Jersey market, Wong said, "You really have to lean on how you infer, how you have much more nuanced inferences from transactions of very different houses."

Despite the complexity, there may be a growing base of knowledge helping Opendoor's machine learning models, indicated Wong.

"What's been interesting is, there is some marginal signal," said Wong. "There is some benefit to, say, knowing trends in Phoenix and Dallas, and applying that in New York and New Jersey. And we have a lot of our models trained on a national scale, so there are some inferences we able to make across markets."

Also: How Opendoor beat Zillow

The company is also able to re-use its operations software. That is important because, in addition to data science, the company has to coordinate home project managers who physically go to a site.

"It's not only about delivering the customer experience; it's also about boots on the ground," said Wong. "We are different from traditional technology companies because we are at the intersection of bits and atoms."

As far as volume of transactions, Wong said, "We can see New York and New Jersey being a major driver of revenue for us." (Opendoor clarified in a follow-up statement that the new market "could be a future driver vs. short-term driver, as we're just launching the market.)

"I'm personally hopeful it will scale fast," said Wong.

Big Data

How to find out if you are involved in a data breach (and what to do next)Fighting bias in AI starts with the dataFair forecast? How 180 meteorologists are delivering 'good enough' weather dataCancer therapies depend on dizzying amounts of data. Here's how it's sorted in the cloud

• How to find out if you are involved in a data breach (and what to do next)
• Fighting bias in AI starts with the data
• Fair forecast? How 180 meteorologists are delivering 'good enough' weather data
• Cancer therapies depend on dizzying amounts of data. Here's how it's sorted in the cloud