Compass’ ‘Zillow Tax’ is another misleading stat

Compass put out a post you may have seen. Black background, big yellow letters:

THE “ZILLOW TAX” — Homes Displayed on Zillow Sold for 1.3% Less.

Then, down at the bottom, in type about one-eighth the size, sits the part almost nobody reads:

“The median sale-to-list price ratio for listings not on Zillow is 100%, compared to 98.7% for listings on Zillow.”

Read those two lines again. They are not the same claim. They are not even close.

The big letters say sellers lost money. The small letters say something completely different — and a lot less exciting.

What they actually measured

“Sale-to-list price ratio” sounds like something you need a calculator for. You don’t. It just means: did the house sell for its asking price?

List a home at $500,000 and sell it for $500,000, and you’re at 100 percent. Sell it for $490,000, and you’re at 98 percent. That’s the whole thing.

So when Compass says its off-Zillow homes came in at 100 percent, and its Zillow homes came in at 98.7 percent, here is what they have actually told you: Our sellers got their full asking price. That’s it. That’s the finding. And here’s the trouble. The asking price is not a fact about the house. It’s a number the agent picked.

The student who wrote his own test

Picture two identical houses sitting side by side. Both are truly worth $1 million.

• Agent A prices hers at $1 million. It sells for $987,000.
• Agent B prices his at $950,000. Three buyers fight over it. It sells for $950,000.

Agent B got 100 percent of his asking price. Agent A only got 98.7 percent. By Compass’ measurement, Agent B is the winner.

Now look at the sellers. Agent A’s seller went home with $987,000. Agent B’s seller went home with $950,000.

Agent B’s seller lost $37,000 — and lands in Compass’ winning column.

That is a student writing his own final exam, taking it and then grading it. Of course he aced it. He picked the questions.

What the honest post would say

Strip out the spin and here is everything the data shows:

Compass homes that weren’t on Zillow sold for 100 percent of their asking price. Homes on Zillow sold for 98.7 percent of their asking price.

Now notice what disappears. There is no “less.” There is no “cost.” There is no “tax.” And there is definitely no $5,590 — the dollar figure Compass came up with by taking that little gap and multiplying it against the price of a typical American home.

You cannot turn “how close did we get to our asking price” into “how much money the seller lost.” Those are two different things. It’s like measuring how tall someone is in pounds.

Nowhere in this study did anybody figure out what these houses were actually worth. And if you don’t know what a house was worth, you cannot say the seller got less than they should have.

Here’s what the headline should be

The number Compass chose is exactly the number a quiet, off-market listing would be expected to win — for reasons that have nothing to do with the seller making more money.

When a house is marketed quietly, one agent sets the price with almost no information coming back. No weekend packed with showings. No flood of calls. No competing offers telling that agent the price might be low.

When you’re guessing in the dark, you guess low. And when you price low, you get your asking price. Sometimes you even get a bidding war. 

So here’s what the headline should have said: Private listings sell for 100 percent asking price because they priced the property too low.

Everybody argued about the wrong thing

What surprised me most is that the people tearing this study apart missed it, too. Zillow’s economist called it unserious. A professor called it a marketing piece. An economics firm said you can’t reasonably compare 806 homes against roughly 296,000.

All fair points. And every one of them is about the size of the groups.

Not one of them said the simple thing out loud: It doesn’t matter how many houses you run through this. You could do it with every home in America and still not know whether one single seller got more or less money.

What to say at the kitchen table

If a seller shows you that post — and some of them will — you don’t need statistics. You need three words: “Less than what?”

Less than the house was worth? Nobody measured that.

Less than it would have sold for out in the open market? Nobody measured that either.

The only thing it is “less” than is a number an agent made up.

Then give your seller the comparison that actually counts: What an outside estimate says the home should bring, next to what it really brought. That’s what Zillow’s research tried to do across millions of sales, and what the Bright MLS findings looked at too.

You can argue with what those studies concluded. You can’t say they measured the wrong thing.

One last thing

I want to be fair here. This is not an argument that private listings are always wrong or that Compass is acting in bad faith. There are sellers for whom privacy is worth real money, and a quiet launch handled honestly is a legitimate service.

This is an argument about evidence. We are suing each other, filing complaints with regulators, and telling buyers and sellers what to do based on studies like this one — and the studies are getting worse, not better, as the stakes go up.

When a brokerage, a portal and a professor can each hand you a number pointing in a different direction, agents lose the ability to tell a client anything with confidence. That’s the real cost, and I’ve argued before that this whole mess started with a policy nobody wanted to re-examine.

Even Zillow’s own economist is fighting this out over who private listings really serve. Fine. But before we argue about the answers, we ought to agree on what the question is measuring.

The fix isn’t more studies. It’s better questions about the ones we already have. And the first question never changes: Less than what?