Big-time real estate advertiser Zillow learns a hard lesson that flaws in its Zestimate home valuation tools translate into huge losses in the iBuyer market. Posting huge losses, Zillow pulls out of the Zillow Offers business and sheds 25% of its workforce as a result. This surprised a lot of industry analysts, but many Real Estate Agents weren’t surprised.
I remember a time not that long ago when we would go on a listing appointment and present our clients with a highly thought-out Competitive Market Analysis (CMA) only to have them pull out the Zillow Zestimate and argue that “according to Zillow” their home should be priced much higher. You never really heard much from the owners where Zillow had their Zestimate significantly lower, but you would hear from buyers, saying a home is overpriced. Sometimes it was a tough battle if you came in below the Zestimate. Zillow after all was touting some very impressive ratios between their on-market home estimates and the final sales prices. But for us, something was amiss and the key was that “on-market” qualifier.
Time after time, our numbers came in spot on, so how could you reconcile those differences? Well at the time Debra and I happened to be pretty heavy in the flipping market, and as it turned out that gave us special insight as to what was going on. When you flip houses, due to the quick turnaround, you tend to buy and sell under the same market conditions. What changes with a flip is the condition of the home. Zillow can’t judge the home’s condition from past sales, tax records, old photos, or Google Maps, so here their errors are amplified. They also can’t judge functional obsolescence, lack of curb appeal, and features not tied to square footage. So how does Zillow end up with such tight ratios? Well, I can tell you from watching them closely on my flip properties, they sort of cheat.
Being suspicious and curious, when I knew I was going to make an offer on a flip, I hurried up and checked the Zestimate and made a note of it. Then a short time after I purchased the flip, I also noted the Zestimate. The Zestimate adjusted to what I paid for it, except usually a little higher. That is not particularly surprising, because that is kind of what you would expect. I am not sure why it tended to be slightly higher than what I paid, but maybe they wanted to make me feel good about my purchase, and maybe grow fond of the Zestimate? Perhaps, but what happened next was more revealing.
I checked Zillow just before I relisted these houses just to see if the value had changed. Usually, it did not. But as soon as my listing would hit the market, Zillow adjusted its price upward to reflect me listing the house at a significantly higher price. It was nowhere close to the higher price I listed because we were high return flippers, but it was in the right direction. Then when these houses would pend, what would Zillow do with their Zestimate? It appears that they substantially move toward the list price, without knowing the actual sales price.
Exactly where Zillow draws its line between off-market and on-market is not clear, but it misses the point. The point is that the Zestimate’s implied value is that it is a good predictor of a home’s worth. The truth is, that in order to judge its value as that type of tool, you would have to compare Zillow’s off-market Zestimate just before a home lists to the final sales price to make that determination. Adjusting it after the listing is on-market is cheating because it is heavily weighted on somebody else’s hard work. It is kind of like an appraiser changing his numbers based on the list price and sales price. Zillow will tell you that they have a sub 2% on-market to sales price error, but if you read further their off-market median error rate is more like 7 to 7.5%.
Historically in the Indianapolis market, the sales price to final list price ratio is about 97%. That has changed due to the housing shortage, but the point is, if houses are listed too high, the tendency is to lower the price until you get within 97% of what a Buyer will offer. Otherwise, they tend not to sell. If Zillow adjusts their Zestimate based on the current list price, which it appears they have done in the past, well they too will have a very good on-market to sales price ratio of at least 97% if the home sells.
Zillow’s flipping failure seems to confirm our suspicions regarding the poor use of the Zestimate. In fact, by some nationally published estimates, Zillow lost nearly $100K per house they flipped, losing nearly $420 million in just 3 months according to the New York Times. To me it appears clear that they were unable to accurately predict the proper purchase price and future sales price of their target homes using automated valuations (Zestimate). And that was during a market where list price was considered the floor of home sale prices. Other problems like trying to scale up a process for flipping homes across the many geographical variations in the labor market had to also be a nightmare.
There are a lot of copycat investors in the iBuyer market, and a lot of Zestimate clones. It remains to be seen how this plays out across the board. There are a lot of institutional buyers in the buy-and-hold market that have pushed rental prices and first-time homebuyer markets to unsustainable limits. At some point when vacancies or defaults increase due to unaffordable rent, those homes will start flooding the market. It seems like that day will be coming sooner than later. The ominous sign is when the average investor is looking for rental properties in an increasingly scarce market.
