Trending: Microsoft AI chief Harry Shum to depart in February after 23 years at the tech giant

valueappeal1The other day, I was talking to my brother about the high property taxes he pays on his Seattle home. Being the good brother, I suggested that he check out ValueAppeal, a nifty online service which we’ve written about several times on GeekWire over the years that essentially allows home owners to dispute their property tax bills (and save hundreds or thousands of dollars in the process).

But, as it turns out, ValueAppeal is heading into the sunset.  I got this note from my brother after he recently tried to use the service: “ValueAppeal says it can no longer evaluate tax appeals.  What’s up?  Looks like a great service to me.  I wonder who shut that down?”

In fact, ValueAppeal CEO Charlie Walsh made the tough decision to shut down the service earlier this year, saying the economics of helping homeowners challenge their tax bills for as little as $99 just didn’t work out.

“There is only a thin slice of homeowners that are overassessed, about 20 percent of homes,” says Walsh. “And at the same time there is only a 30 to 60 day window in which you are allowed to file an appeal, so that presented a marketing challenge.”

mostlikely1111In order to solve that challenge, ValueAppeal turned to direct mail. “To make money at $99 a pop acquiring customers through direct mail — you have to be pretty good at direct mail,” said Walsh, adding that they started layering on all sorts of demographic data about who was most likely to respond to direct mail campaigns and from that who was most likely to file a tax appeal.

“Over the course of a few years, we got really good at predicting who would respond to our direct mail,” said Walsh, describing the data as a “crystal ball.” Late last year, Walsh started kicking around new ideas, thinking about ways they could predict which homes would be most likely to sell, refinance or foreclose in a certain geographic area.

Charlie Walsh
Charlie Walsh

They started testing a new service called MostLikely a few months ago with real estate agents and brokers. The response was fantastic, and Walsh and his team decided they needed to pivot the business.

“Making a decision to change your business model should be a metrics-only business decision and not an emotional decision, but of course it is our baby and we’ve been building it the last couple of years, and so it was an important decision and a big decision,” says Walsh of deciding to pull the plug on ValueAppeal. “It was the right decision, and we have refocused our efforts.”

With the change in direction, ValueAppeal is no longer accepting new customers. Walsh says about 95 percent of the technology transferred to MostLikely, forming the basis of that new product.

“We are just now asking different questions of the data,” says Walsh.

Instead of trying to help homeowners figure out if their tax bills are too high, MostLikely is attempting to help real estate professionals develop leads through predictive intelligence, a business that’s very much of interest to companies like Trulia (which just bought Kirkland’s MarketLeader) and Zillow.

mostlikely44Walsh admits that the “pain point” still very much exists for home owners trying to appeal property taxes, but he said the company could not afford to service those customers while going in the new direction. “Any startup advisor will tell you: ‘Focus. Focus. Focus.’ So, trying to support our property tax appeal service, while growing a new business is beyond the resources that we have,” he says.

Earlier this year, ValueAppeal scooped up $6 million in additional financing, and at the time Walsh said the cash would be used to roll out a new product. That the new direction is MostLikely, with several products directed at real estate professionals.

The first of which is dubbed, which as the name suggests attempts to use predictive analytics to tell real estate professionals which homes have the most likelihood to be listed for sale in a specific area in the next six to 12 months. Walsh dubs it “MoneyBall” for real estate agents, referencing the classic big data tactics that Oakland A’s general manager Billy Beane used to build that franchise.

Though data varies by geography, MostLikely says that it can identify the properties that are between 1.5 times to 5 times more likely than average to sell in six to twelve months. That can help real estate agents more intelligently spend their marketing dollars, he says.

Walsh and his team are packaging those leads into a six to 12-month subscription service, which allows agents to pay $400 to $500 per month to get data in a set territory of about 2,000 to 3,000 homes. Agents get an exclusive on leads in that area for the subscription term, which is similar to the old model (which morphed into Market Leader).

“We are providing a competitive advantage to our real estate agent clients to reach out and touch these homeowners who are most likely to sell their home before any of the competitors do,” he said. “It is a pretty compelling offer, especially in this current market.”

The company is also working on predictive models around refinancing (; reverse mortgages (MostLikelyToReverseMortgage) and foreclosures (

Follow us on Twitter @GeekWire.

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline


Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.