This is a fun little passion project by Airbnb-engineer Andrew Pariser and someone known as Potch. The way it works is that it shows you a picture of a recently sold property, and you have to guess what it sold for. You get a bunch of guesses, and after each one, you are given more information about the property and some feedback on how close you are. To win, you need to get within 1%.
When I tried it out, my initial guess was way off (too high). Toronto has trained me well. I also wasn't sure where Evansville, Indiana was, so that bit of information didn't really help me. But the arrows telling me I was way too high, certainly did. The reality is that it's pretty hard to guess the value of a home if you don't know where it is, you can't see interior photos, and you generally don't have enough information.
But what if you were from Evansville, Indiana and what if you did have enough information? I bet that the guestimates would actually be pretty accurate. This idea of crowd-sourcing market information and pulling wisdom from crowds has long interested me, because price discovery is a major pain point for real estate. Sure you can look at comparable sales and current listings, but that is not an exact science. Neither are algorithms.
But what if there was a way to test the market and get pricing feedback before you actually list? Would you trust it more than Zillow's algorithm? This is something that I'm working on testing right now through a passion project called Unlyst. Myself and a few others are working on a very simple product that will be released this fall. If you'd like to follow along, sign up here.
