TopHouse Home Search App Outperforms HouseSigma and Zolo Home Valuation Estimates
December 7, 2022
TopHouse recently launched a home valuation estimates and we wanted to run an experiment between some of the other popular home search apps to see which valuation algorithm is better (HouseSigma and Zolo). We were surprised with the results.
Here is how we ran the experiment.
We randomly selected over 300 homes that were sold conditionally on various days in the last month. We recorded each apps valuation estimate before the sold price was recorded. We selected the first to report their price.
While some people know that automated home valuations are not to be trusted, we've noticed that a large number of home buyers rely on these estimates in determining what they bid on a home. This can be a huge problem if the estimates are off. We recently ran an experiment where we noticed 80% of estimates by HouseSigma and Zolo are over valued 🤔
TopHouse recently launched home valuation estimates and we ran an experiment between some of the other popular home search apps to see which valuation algorithm is better. We were surprised that TopHouse was usually the closest estimate to the actual sold price with the least overestimation bias.
Here's how we ran the experiment:
We randomly selected over 300 homes that were sold conditionally on various days in the last month. We recorded each app's valuation estimate before the sold price was released. It was important to capture a snapshot, since estimates naturally change over time. As soon as a listing sold, we marked it in our spreadsheet and compared the actual sold price to the estimate.
Here are the results.
Example of TopHouse Valuation: https://tophouse.com/l/417JoY
TopHouse was the closest estimate for 52.78% of the test sample with an average difference of 4.83%. HouseSigma and Zolo’s accuracy were 6.61% and 8.99% comparatively.
The most surprising discovery was that in approximately 80% of cases, HouseSigma and Zolo’s estimate had overestimated property values (ideally this would be around 50%). While this might happen in a down market, a smart valuation algorithm should pick up on this 6 months after a downturn.
Feel free to double check on the TopHouse app and run your own experiment.
The results of the test are in the spreadsheet below: https://docs.google.com/spreadsheets/d/1hY1QAi6MWd7XN2A_7DrDbcrGaDn8W0k8k5TK8y8ssXc/edit?usp=share_link
Screenshot samples: https://drive.google.com/drive/folders/1fuHzPws3dCeV_U5VmTid2MgZhPeF_iF3?usp=share_link
Let us know how the TopHouse app stacks up. We’d love to hear your thoughts and feedback.