Despite having reported on it for several years now, there are still plenty of times when I don’t understand the real estate industry. One of them came again this afternoon, reading an article by the “Demon of Marketing” on Inman.
With the news of the Zillow/Trulia merger finally getting green-lit this week, there’s been tons and tons of speculation on What It All Means for real estate, and especially how the landscape will change with Zillow and NewsCorp competing head to head. The Demon’s take is that it will all come down to which of the portal owners is able to launch an “innovation bomb” that takes out the other. Specifically, he thinks that bomb will come in the form of an AVM, that is, an automated valuation model.
AVMs already exist, of course. Appraisal companies and banks have used them for decades, especially for refis (much to experienced appraiser’s dismay). There’s even been a move by some brokers to get permission to develop their own AVM models, as a marketing tool or a tertiary business. There’s also a fairly well known web tool which is, at heart, an AVM aimed at giving consumers an easy way to find out what their house is worth: It’s called a Zestimate and real estate agents hate it with the white-hot passion of a thousand suns.
Zillow introduced the Zestimate about 10 years ago, in fact. But, the Demon says, the reason that this new AVM he envisions will be paradigm-shiftingly, industry-upendingly innovative is that it will also fully incorporate the “emotional triggers” that truly drive real estate purchases. Current AVMs are based purely on simple, numerical data about the house – square footage, tax assessments, number of bedrooms, recent nearby sales. The kind of AVM he envisions will be much more than that, it will, someway somehow, account for “flow,” finishes, feel, the kinds of ineffable properties that make a place feel like a home instead of a mere building.
Right. So, the thing is, the reason agents hate hate hate Zestimates is because they think they’re inaccurate – Zillow itself admits (if you take a deep dive into their fine print) that in the majority of cases the Zestimate is merely within 8 percent of the sale price, and that in many towns, many neighborhoods, for many houses, it may be way farther off than that. And, as mentioned, the Zestimate is derived only from hard, factual data and sophisticated but well-understood math – regression analysis, etc. And yet Realtors on the whole find inaccurate Zestimates which give homeowners a false impression of what their house is such a huge problem it colors their whole opinion of the company.
So to me, reading the Demon’s piece – his inevitable “innovation bomb” feels like the Susie’s Pony of Internet real estate. I mean, sure, as long as we’re dreaming, why not ask for a pony? Why not imagine a perfect automated system that allows a seller to know exactly what their house is worth, or a buyer to find the exact right property for them? (Though one might wonder what the agent’s for, then, in this dream world. If we don’t need them to help find the house or negotiate the sales price, 6 percent seems rather a hefty fee for taking an afternoon off to wait for the building inspector.)
It’s true, buying and selling real estate is definitely more Kirk’s game than Spock’s – knowing a property’s square footage tells you very little about the things that make one home more appealing than another. But real though they are, the emotional ties that turn a particular property into a dream home are incredibly hard to pin down – and quite likely to vary a great deal from person to person. It seems impossible to me that anyone could algorithmically generate a more accurate estimate by incorporating even less precise inputs.
Yet, and here’s the part I don’t understand, the comments on this piece think it’s genius. I mean, yes, I guess it would be great if we all had ponies.
