‘Zillow-killer’ Blossor out to conquer natural language search

Listing site 'gamifies' search experience

Don't miss Hacker Connect SF
Take a deep real estate technology dive, Aug 7, 2017

It takes some real “cojones” — perhaps even a touch of hubris — to think you can take on Zillow at this point. The listing portal is now worth more than $5 billion, and it’s dropping $65 million on advertising this year.

But Ben Kinney, a Bellingham, Washington-based real estate entrepreneur, thinks his startup has got what it takes. And based on what partners at Google Ventures have said about “Blossor,” perhaps we shouldn’t rule him out.

They dubbed the fledgling listing portal a “Zillow-killer.”

Kinney’s credentials also inspire some confidence: He said he has come to own seven brokerages with 800 agents after 10 years in the business, has sold a real estate tech startup to Market Leader (now owned by Trulia) and has seen early success with another, the transparent task manager Brivity.

Blossor’s game plan for listing-portal domination hinges on three strategies: conquering natural-language search, insulating consumers from agents, and “gamifying” its user experience. 

The site is currently offering early access on a first-come, first-serve basis, but has plans to launch out of beta sometime this summer.

“We want to create a place for consumers where we can nurture them over time until they’re ready for an actual transaction,” said Kinney.

First, you have to reel them in. Kinney plans on doing that by optimizing Blossor’s rankings in “long-tail” search, or searches that use natural language, rather than key phrases.

Kinney doesn’t aspire to rank high for searches like “Seattle homes for sale.” What he’s shooting for are much more specific search inquiries like, “Seattle condos close to the highest-rated bars” or “Seattle homes over 1 million with a pool.”

“Long-tail search is not something that the major portals have focused [on],” Kinney said. “We want to own the long-tail first.”

The site has dumped “every word in the English dictionary” into its database in pursuit of that goal, and generates a URL for every search that a user types into the site’s search box.

Blossor’s database currently has about 90 percent of listings in Oregon and Washington sourced from MLSs, according to Kinney. By the year’s end, he wants the site to cover the “entire Western part of the U.S.”

Natural-language search may seem to jibe with real estate search, but Marc Davison, a partner at real estate consultancy 1000watt, notes that listing sites have already experimented with the technology for years.

“There are just so many variables to a home search that I believe users will miss out on properties based on their requests not matching up with the details the agents plugs in,” he said. “This feels gimmicky.” operator Move Inc. rolled out a natural-language search tool for real estate professionals, “Find,” in 2010, which it offers to MLSs. MLSs that have signed up to provide Find for members include Midwest Real Estate Data (MRED) in Chicago, MLS Property Information Network Inc. (MLS PIN) in Boston, and the San Francisco Association of Realtors.

Move recently revealed that it’s invested in a startup, Ylopo Inc., that’s out to build “the world’s most intuitive (real estate) search engine.”

Kinney thinks Blossor’s natural-language search blows away everything that’s come before it.

Blossor custom view showing thumbnails of homes by price -- in this case, homes in Seattle priced above $800,000.

Blossor can show thumbnails of homes by price — in this case, homes in Seattle priced above $800,000.

“Real estate sites and MLSs might do simple keyword searches like ‘with pool,” he said. “Blossor looks at MLS data, public data, and social and online databases to allow users to search in sentence.”

Blossor is also working constantly to improve its search tool, he said. Every time a user’s search turns up zero results, Blossor’s team tweaks the site’s algorithms so the tool supports the search the next time around, according to Kinney. 

Blossor’s beta testers have conducted over 40,000 natural-language searches that have helped Blossor support queries like: “I’m looking for a home in Seattle with a gourmet kitchen,” “I need a home in Bellingham where I can have horses,” and “Bellingham homes for whale watching.”