Zillow’s algorithm is fundamentally flawed. Since they the goal is to monetize Zillow and since consumers think the estimates have validity, this is one heck of a problem.
We’ve been tracking some bizarre Zestimates and the Internet is awash in complaints that Zillow estimates that are too high, too low, or dropping/rising too quickly.
Take your pick!
Zillow flaks (currently David G.) are quick to show up when a new post appears criticizing Zillow, but much slower to proffer any reasonable explanations of the discrepancies. Indeed, they’ve yet to offer anything even remotely resembling a reasonable explanation.
The problem, in my opinion, is not that the model is flawed, but that the algorithm is a black box and consumers use these data as significant inputs for decision making.
The problem is compounded when Zillow justifies the errors by in effect saying, what’s the big deal? people don’t use these numbers anyway.
Sorry, but a profit-making concern can’t have it both ways.
Zillow needs to get better fast or suffer the consequences.
See this thread for a brief discussion of what I consider to be astonishing errors for three model matches very near each other and Zillow’s lame explanation.
