This analysis was first published in SvD Näringsliv, in Swedish, on November 8th, 2021.

Using algorithms, the American real-estate giant Zillow was going to buy homes and add $20 billion to annual revenue. Instead, the project is now being shut down and thousands of people laid off — all because the software miscalculated.

“Good afternoon, everyone. We appreciate you joining us today.”

It sounds friendly, but it was probably mostly something he had to say. Rich Barton, CEO of Zillow, sounded tired and deflated when he delivered the bad news on his earnings call last week.

The company, which can be described as an American Hemnet, announced that it is winding down its Zillow Offers division, taking a $569 million write-down (about SEK 4.8 billion) and laying off nearly 25 percent of staff. After riding high on the wave of rising US home prices, the share fell hard – more than 30 percent in just a few days. How could this have gone so wrong?

The phenomenon is called “iBuying”, short for “instant buying”. The term was coined by an equity analyst in 2017 to describe a new type of company in the housing market. Services like Opendoor and Offerpad offered a fast and easy way to sell the most expensive – and often hardest to sell – asset most people own: their home.

iBuyers develop algorithms that compute what a home is worth and produce an offer within minutes after the seller has filled in some basic information. Accept the offer and the sale can close within a few days. The business model is built on cutting out the broker, and the companies therefore buy the homes at a discount compared with what a bidding war could have produced. Once you subtract broker fees and account for the speed, the idea is that the offer is roughly equivalent.

The model exists in Sweden too. The company Movesta recently raised SEK 85 million in venture capital, and in an interview with SvD Näringsliv in March, Micha Lindqvist, one of its founders, said they receive hundreds of applications a week.

It sounds easy and clever. But as Zillow has just learned, the model has its limits. As anyone who has moved between cities knows, the housing markets within a country can vary dramatically. Zillow had therefore focused on certain regions and tuned its systems to perform well precisely in the areas it served. But as with so much else, a company’s algorithms are only as good as the way they’re programmed.

In this case, Zillow had pushed its bids up just as the housing market was starting to cool. Suddenly the company was sitting on 7,000 homes that wouldn’t sell, and had to offer them to institutional investors instead. “I think they relied on home price appreciation at exactly the wrong time”, as analyst Ed Yruma told Bloomberg.

There are two ways to read the failure.

You can see it as an example of a disruptive company that promises the moon but ultimately can’t deliver. We’ve seen it many times before – from the infamous American e-commerce company Pets.com to the Swedish group-buying service Letsbuyit. Companies that present a compelling vision of the future but in the end can’t live up to it. Maybe not because the vision itself was wrong, but because the timing or the execution didn’t make it the whole way. Selling homes without a broker is a radical change. But maybe brokers offer something the algorithms still can’t.

The other way to read Zillow’s failure is that big companies have a hard time pivoting their business model. Opendoor, one of the first iBuyers, spent all its time getting the pricing and selection right – it picked a niche, a single track, and aimed to be the best at it. Zillow – the American Hemnet, as noted – already dominated in the US with many other services when it decided to add iBuying to the mix. But it’s hard to retool around innovative business models, especially when it’s only one of many things you’re doing. The smaller companies’ speed and focus can sometimes tip the scales, even with fewer resources.

Is this then the end of iBuying as a phenomenon? The stock market doesn’t seem to think so. Competitor Opendoor’s share price did fall when the first news about Zillow trickled out. But once it became clear that Zillow was giving up on iBuying – Opendoor rose 19 percent.