One day, you notice a nearby home has a “for sale” sign attached to it. “I knew it!” you say to yourself. It was only a matter of time. The neighborhood is hot and owners are rushing to cash out. With a new shopping center and office park getting built down the road, the area is growing. Also, with many homes approaching 20 years old, many “empty nesters” — with the kids grown and gone — are ready to downsize to something smaller and are putting their homes on the market.
We developed a model to determine whether a house will be listed for sale the near future based on factors like these.
Ways you can use it
If you are a real estate agent or a bulk buyer and want to send flyers to homeowners that have a higher chance of selling their homes, you will be much more efficient with your marketing dollars by targeting homes that have a higher chance of selling in the near future.
Similarly, if you are building a product or service that depends on your homeowners selling (or not selling) in the near future, perhaps because you need to recover costs in a certain period of time, you also want to have a systematic way of determining the likelihood of a home selling.
What determines probability of a home selling
Local market conditions
The local environment differs significantly by neighborhood in terms of supply, demand, and performance. Is the neighborhood going through an upswing in prices? Is the area outperforming (or underperforming) the metro area? Is there a flurry of renovation activity or flips happening in the area that might motivate a sale? We look at all these factors and many more.
The home’s specific situation
Individual characteristics of a home also make a difference: Are mid-priced 3-bedroom homes hot on the market right now? If so, similar homes might be listed sooner. How long has the home been owned? Is it renter-occupied? These factors matter too.
Putting our forecasts to the test
We decided to see how well our analysis works. We looked at homes sold from mid-2020 to mid-2021 and built a model that forecasts the probability of other homes also being sold in the near future. We did this on a city with 30,000 single-family homes.
In the map below, we show homes that sold in the last year in black. The rest of the homes are color coded according to the probability of a sale. Red-shaded homes have the highest odds of being listed, exceeding a 29% chance. Orange shaded homes have a “moderate” likelihood. Lastly, light-green shaded homes have odds that are slightly elevated.
5x your marketing ROI
If you randomly chose 500 houses in the same city of 30,000 homes, about 4 to 7% of those homes sold in the last year.
Meanwhile, if you had marketed your services to the top 500 houses in the same area based on their odds of a sale, 36% of those houses actually sold. You’ve just improved your odds by 3 to 7 times, along with your marketing ROI.
Put another way, you only have to make one sixth as many mailings to get to a sale.
