Hacker News new | comments | ask | show | jobs | submit login
[dupe] Zillow surprises investors by buying up homes (techcrunch.com)
76 points by BukhariH 9 months ago | hide | past | web | favorite | 90 comments

Buying homes for subsequent resale is a capital-intensive, old-economy, bricks-and-mortar business.

Zillow will borrow money to buy the homes, which means that (a) the clock will start ticking the instant each new home is purchased, and (b) this endeavor can be profitable only if proceeds from resales/rentals are sufficiently high to cover cumulative debt service costs -- in addition to all property taxes and ongoing maintenance expenditures associated with home ownership.

Why would a heretofore capital-light SaaS business like Zillow want to do this?

The only sensible explanation I can think of is that Zillow's current business is no longer growing quickly, i.e., Zillow is now a boring, mature company.

The stock dropped 7% on the news last Friday.

Or Zillow has access or insight to data that general property investors do not, thus driving abnormal returns.

That's my guess also. If they have details on the specific aspects in the specific markets that drives prices, being able to "jump on a deal" could work out well.

The question becomes if there are enough opportunities like that to move the needle for them.

If I had to guess, I would have guessed moving into Lending Tree's matchmaking area would be easier and have a better ROI.

What specific details would they have that other's don't. I don't know how prevalent FSBO is on Zillow in some markets, but in mine it's extremely limited. I'd be shocked if it made up 5% of the residential market, probably closer to 1%.

The rest of the listings are from MLS systems and they pull housing data from county systems. For that, I'd assume there is a data broker and Zillow isn't integrating with thousands of counties. If that's the case, then both the MLS and county data can be had by anyone. With that information, you'd get lot details, house details, listing history, county tax valuations and sales history.

What other data do you need?

A lot of house flippers are agents because they get the MLS listings ahead of the public and can move faster on them. I don't know when Zillow gets their MLS data, but listings almost always hit Realtor.com a day before they hit Zillow, so I don't see that as an advantage either.

> The rest of the listings are from MLS systems and they pull housing data from county systems. For that, I'd assume there is a data broker and Zillow isn't integrating with thousands of counties.

That's not the case, MLS is a giant mess and the reason why Zillow was/is amazing is that it aggregated and normalized lots of different unconnected systems. They still have gaps and it's common for an individual MLS to play hard ball and not give out their data, but it's still a large competitive advantage.

What's FSOB? Was it just a typo of FSBO (as in for sale by owner)?

Yes, corrected.

They even know unlisted data, like which buyers are currently looking. If they combine it with data from social network partners and other providers they can know the school-friend networks of the buyer's kids, church affiliations, how racist they are, etc.

Social networks don't sell data like that

Wouldn't a place like Zillow have first hand knowledge? You can sign in with your facebook account, for example. Linking your account will give them your public profile, friend list and email address.

Do they have to sell the data for Zillow to figure it out? I assume at a certain point you get enough data from enough sources to make those inferences yourself.

You can sign up to Zillow via Facebook, where they get permission to pull the most relevant data points. They can then make additional inferences based on that.

What do you think are the most relevant data points that you believe can be pulled from that API?

Among other things they get access to "age range, gender, language", which in combination can often already be a good starting point. They also get access to the email address which can be used to pull more info via other APIs, if you have used that email much in the past. They also get access to "other public info", which I'm not 100% sure about, but I think that should be equivalent to what you see when you go on someones public profile page, and includes Page likes for a lot of people.

Facebook offers targeting. They don't sell "data points".

"Sign up via Facebook" != "Facebook Ads"

Or Zillow simply has a lot of data and can scale it up.

I suspect part of the reason is that SV "unicorn" Opendoor (https://www.opendoor.com/) is doing this and it wants a piece of the pie.

Question I have is if whatever "pie" Opendoor is grabbing is actually sustainable in a recession/down market...seems like the worse case scenario is that they are left holding a bunch of illiquid inventory of declining value they can't get rid of.

The risk is in sitting on a lot of properties and getting overextended during a downturn. I think that can be mitigated because Zillow/Opendoor's model is not to buy and hold, but to:

1. buy at a discount (essentially a fee for the service of selling your house with a click)

2. sell at a markup (we're talking a small one in most cases, but the fact is in a good market you can reliably print small amounts of money with some paint, some minimal landscaping, and new kitchen appliances. Zillow will not be using a high-interest loan for the purchase or repair, so they will not be sweating like your typical flipper on TV.)

Assume for the sake of argument that Zillow has data to decide where these bets are safest based on comparables and key economic indicators. And they will say no to sellers as often as the data suggests they should.

I don't think this has much to do with finding super-profitable deals with data, but with reliably shaving points off a large pipeline of deals.

If it doesn't work in a healthy economy, they stand to lose the difference between the discounted price they paid for a house and the market value they can sell it for. In a recession, they can probably stay afloat by renting properties they can't sell at a decent price.

With regards to your point about expensive financing, this moves Zillow from tech company to REIT; a dangerous shift in a rising interest rate environment depending on how long they end up having to hold inventory for.

We are very far along in the current economic cycle, and a downturn is inevitable. It would be unwise to be sitting on a large real estate portfolio you can’t quickly unload when that occurs.

Will be interesting to see if this Hail Mary pays off.

It is interesting they even choose the same cities to start in.

Vegas makes sense...the best markets for something like this are ones where there's a good amount of inventory, median prices are not very high, avg time on market is on higher side (giving a reason for sellers to use Opendoor to trade value for liquidity) and there's a good balance between buyers/sellers.

Improving liquidity in the market, aka being a market maker, can provide an interesting advantage.

Agreed. The surprisingly bespoke nature of home translations definitely provides an opportunity for this to be a win/win.

Zillow wins by buying an asset it has more information on the true market value of, then providing an easier, economies of scale on the associated services (inspection, legal, repair, etc), then being able to offer a standardized product to home buyers.

Buyers win from taking a lot of the uncertainty out of a transaction. And from possibly lower fees if Zillow decides to rebate a portion.

Let's not forget there's a standard 6% commission in US real estate transactions... most retail would kill to start with a 3% margin.

If Zillow has high confidence a home is underpriced, it makes sense to time arb that into profit when it can match with a buyer.

Let's not forget there's a standard 6% commission in US real estate transactions

I never really understood this, the last time I sold a house in the UK the fee was 1.5% and only on my side, the buyer pays nothing.

1) The 6% commission is negotiable.

2) In the UK, the buyer pays stamp duty (tax) on the purchase.[0]

3) You can put a bid on a property in the UK and it's non-binding (until exchange)[1]

In other words - very different markets with very different dynamics.

[0] - I forget the percent, but it's material. This effectively disincentives people from buying, making the market less liquid.

[1] - This is frankly ridiculous.

(I've bought and sold in both US&UK markets)

Did you know that "realtor" is a registered trademark?

What is the generic mark? Real estate agent?

You don't need to pay it, but people feel that it is worth it/ You can definitely get a broker who works by the hour for you.

or flat fee - which will crush every realtor going fwd.

Welcome to the US.

Real Estate is one of the older investment games in town - It is what I am basing my retirement on.

I wonder how this will impact disclosure rules versus sales advertising? Existing model to use, or adopt another industry's? Will this cause fears of LIBOR-style manipulation?

This will be interesting to watch...

edit: p.s. Maybe another chance to make a Carfax-like system for homes? In the US market, I find there is not a lot of opacity, especially when one gets into larger/commercial deals.

What do you envision for this Carfax-like system for real estate?

What info would be available that you could not see via Zillow or the MLS listing?

I think you mean not a lot of transparency.

All businesses can be framed as “old economy” but cheaper. Lyft is just “taxis” done better. An algorithmic hedge fund is just old-style investing automated.

What's peculiar is that they don't use OPM for this. There are literally a million of these flippers out there. Why not just work with them to buy it and sell it, and Zillow gets a percentage for providing the lead? It's like zero risk for Zillow, all upside.

They must be making a lot of money from this if they're going it alone. The 7% drop is probably not because of riks here but because it signals lack of growth.

Because they think they have some competitive advantage in the space?

How can zestimate be trusted if zillow is in the business of buying and selling homes? There's a reason why banks separate advisory business from the trading business.

It's already pretty inaccurate today (overestimates by about 10%[0]), so there's not much change there. ;-)

[0]: Very rough estimate, may be market-specific.

Really it's only inaccurate because the price of housing isn't static and certainly isn't objective. Zillow works based on how similar houses near yours have sold, which could be low because the seller wasn't driving a hard price and the buyer was a good negotiator. It could be higher because the seller was firm on the price and the buyer didn't know any better. Not to mention, Zillow doesn't know if you recently renovated your kitchen. They don't know if you have unrepaired water damage. It's all a guess.

That being said, I wouldn't say 10% is a super high overestimation. On my street are two identical houses and one sold for $150k while the other sold for $180k, even though they're literally identical other than the color of the siding.

The number that really matters is the appraised value. Zillow had my house at $110k when I bought it and the buyers were asking $140k. It was a surprise when the appraisal came back at $145k. If Zillow was right, the bank would never have given a loan for $30k over the value of the house. But there was no way for Zillow to know the amount of work the previous owners put into the interior of the house.

> Really it's only inaccurate because the price of housing isn't static and certainly isn't objective.

If it was static, an estimation tool wouldn't be useful.

And it's not just "the price of housing isn't static and certainly isn't objective." After all, Redfin faces exactly the same difficulties. Somehow Redfin uses more or less the same data and comes up with estimates that are much closer to the prices houses actually sell for (in this area).

The median house sold in my area cost ~$660k last year (zip code 98117). 10% overestimation is a (much) larger absolute error than $15-30k.

And there is plenty of sale volume in this area of similar houses -- the average/median of which are far below Zillow's estimates.

I wouldn't say appraised value is what matters. What matters at the end of the day is what buyers are willing and able to pay. Appraised value is both a factor in that as well as a result of that.

lol as if banks are actually able to avoid conflicts of interest.

I get the feeling a company like this would only make this kind of leap because they have the data to show it will pay off. I.e. they can formulate algorithms to make the best purchases and the most profit. I foresee big gains from this

It’s rarely a good idea to compete against your customers

In this case it seems the customers are unlikely to revolt. Most won't even notice, right?

Great point. For now, it is limited to AZ and NV and for wholesale purchases only. But it seems like a conflict of interest that's hard to prove isn't being abused.

Isn't Amazon's marketplace similar in that regard? They compete with other sellers using Amazon's platform all the time.

Ethically, it seems a bit fuzzy for me given that one of Zillow's big features is there zEstimate. How can they demonstrably prove they aren't tweaking that somehow for their own gain?

I'd argue that Amazon's customers are the buyers of the products, not the sellers using Amazon's platform.

I'd argue that Amazon has many different groups of customers.

Why? BlackRock offers solutions to other asset management firms to make investment decisions. They use the same solution in their asset management division.

Isn't there a firewall between the units?

Zillow has captured the user base so effectively that the customers will have no choice but to continue paying.

See also: Amazon

Every big company does this in some way. The only smart move in business is to be where the money is.

A testable prediction at least.

I wouldn't be surprised if they make a bunch of money and then make a big mistake and lose a much bigger bunch of money.

This was my thinking as well, they can decide what they want to buy based on their data on prior listings and closing prices. It is an interesting advantage.

But how is this scalable? You need manual labor. This is a bigger headache than hiring agents as full time employees like Redfin.

Zillow has such reach that they can low-ball every offer so that 99% of people won't take it - but as long as 1% of people do, and Zillow can avoid terrible houses, then Zillow will be buying properties at a huge discount from the market, plus collecting a substantial fee. Thus, they can sell at market rates and make a big profit per house.

The key advantage here is that they can make these low offers to a much bigger audience than anyone else can do. This should allow them to be either be more profitable per house than anyone else, do more volume than anyone else, or hit any mix of these two better than anyone else.

Of course, they can shoot themselves in the foot pretty well if they:

- Try to go for volume over profitability, and then catch a downturn. - Do a bad job of running repairs. - Don't do a good job of catching houses that are much worse than they appear.

However, there's no physical reason this can't be extremely profitable. They have the data to see their current home investors making money. If they feel they can identify the most profitable attributes of these flips, then they can route all the extra profitable ones to themselves. The only losers here are the existing people in Zillows home flipping program. They are almost guaranteed to now be getting the second best homes, once Zillow has skimmed off the profitable ones.

Why are they in a unique position with respect to reach? I would think that it would be hard to scale across so many jurisdictions when you need to hire licensed contractors to do repairs and have lawyers allowed to practice real estate law across many states.

Maybe that's not a big part of the cost?

> Zillow has such reach that they can low-ball every offer so that 99% of people won't take it - but as long as 1% of people do

It'd be pretty inconsistent to put in a lowball offer when your company publishes an estimated value for every home in the market.

And yet Zillow has been already been doing these low offers successfully, but with only third parties making the lowball offers.

Sometimes people just want sell now.

So what does it mean for me if I buy a home straight from Zillow? Better price?

Likely you'd get the same deal as any other real estate developer in the flipping business would give you. They'd be looking for undervalued properties that can be fixed up and resold at a profit. Which means they're on the open market and you're bidding against other purchasers, like normal.

The problem seems to be that any homeowner getting an offer from Zillow would immediately know they had mispriced their listing and not accept.

Unless Zillow has some new efficiency that no one else has (super contractors that can fix up cheaper than other investors, low cost capital) I don’t know how they will be able to outbid the market consistently as there aren’t enough dumb sellers with dumb agents out there.

But maybe I’m missing a way how they could use analysis to identify “bargains” without the seller getting clued in and upping their price.

I mentioned this elsewhere, but 'closing certainty' is a real concern in the real estate market, to the point where many sellers will take a lower offer price in favor of a cash buyer.

I never understood why though, in the worst case doesn’t their house just take a bit longer to sell if a deal falls through? Aren’t there backup offers in play? Is taking a lower offer really better than just waiting a bit more for the higher offer?

As an example (real-world) a seller turned down a few early under-list-price cash offers, and eventually got an at-list offer from a mortgage buyer. But the buyer ended up having trouble finalizing the loan and as a result it took almost two months to close. The home had already been sitting on the market for over a month, so had the deal fallen through, the seller may have had to put it back on the market, which could have added a few more weeks at best, and a few more months at worst to the process. So had they gone with the cash buyer right away, they would have sold the home in a few weeks after listing it. But instead it took a few months, and there was a non-insignificant chance that it would have taken a few more. And during this time, the seller was not living in the home, so presumably they were paying rent/mortgage on two homes. For some people that could be tough to swing, and not worth an extra 5-10% on sales price.

A few other things to consider. 1: The above example was in a relatively hot market. But in a more 'normal' market, homes can sit on the market for months at a time before attracting a good offer. So in a lot of markets, waiting for the next offer could mean months at a time. 2: There is a bit of a 'stigmatization' around properties that fall out of contract, because a common contingency is an inspection contingency, so the theory is that buyers start to shy away if a home falls out of contract because they are worried it is a lemon. I have personally seen a few listings where the realtor will explain the contract failure in the top of the post (i.e. buyer couldn't get a loan or something like that) as a clear attempt to ward off the 'lemon' concerns. And because mortgage buyers usually require a long list of contingencies, the risk of a failed contract is a lot higher than cash buyers. So cash buyers are 'safer' in that sense. 3: Sellers are often rolling over the proceeds of a sale into their next house, so any extra money on their home sale will most likely show up as a slightly lower monthly payment on the next house. So for every extra $10k you will only see about $50 lower on your next monthly mortgage payment (assuming 30 year).

Here’s an example reason why someone might take an earlier offer: A new job starts July Nth. It’s now mid-April. The family needs to be moved before July. They don’t have a lot of months left to close a deal.

But how often is a seller so motivated?

Could Zillow use some combination of data from data brokers to find out how motivated a seller really is and then take them for as low as they can go?

For instance you’re selling a home and I see your job has changed on LinkedIn and your home hasn’t sold yet.

It’s about risk mitigation.

There is value in certainty.

It’s why a $100 now is worth more than a pronise to pay in the future

A low priced home isn’t always a mispriced listing. It could be a fucked up home that’s super beat up and no one wants to touch. Or a seller that wants to get out immediately.

I meant low priced relative to value. In your example, Zillow wouldn’t want it either because it wouldn’t turn a profit. Unless they had some super efficient repair crew.

I know nothing about the housing market but to me it seems like this could be a pretty good move from zillow. They could theoretically use data on their millions of houses to find absurdly good deals and only snap those up.

It would be worth it for convenient financing options/on click, even if the price isn’t lower

No advantage though for a cash buyer?

Right, but if the point is to increase liquidity, the low hanging fruit is mortgage backed purchases. Homeowners will often take a smaller offer price in favor of cash because the closing process is easier/more certain. So Zillow might be trying to arbitrage that price difference between cash and mortgage purchases. And if they can find a way to scoop up part of the loan origination fees and the broker fees in the process, than I guess there could be some compelling revenues there.

I'm betting you won't have to pay a commission to an agent

Which should mean a reduced price of at least 3%, since it’s the seller who pays the commission and factors it into the home price.

If you read the Redfin 10k, they were doing this already as a service called "Redfin Now." I think it is interesting that two large data-driven pricing services are both pursuing this strategy. If you have better data, it might be possible to make money just on your pricing edge.

Isn’t this announcement to mess with Opendoor’s current fundraise? Similar to Microsoft announcing vaporware back in the day or Uber trying to interfere with Lyft’s fundraises?

Nobody wants to tackle the elephant in the room in real estate: we need to build a lot more. that's really hard.

Competition with its own customers too. Isn’t this a conflict of interest?

As long as this is disclosed, it should be okay.


This move is very interesting.

I have been studying this part of the business quite a lot lately. A few considerations:

1) It seems to me that Opendoor, Unison, etc, and now Zillow, are trying to capture "pre-foreclosure" opportunities before they hit the market. Think about this: why would you want to sell your house quickly, and leaving money on the table, if not because you're short on cash and you know you will soon lose your house?

2) Secondarily, especially in the case of Opendoor, some homeowners might want to simply avoid the complex and time-consuming task of "changing homes" (fixing small details, picking the agent and listing it, staging it, handling the delicate balance between selling the current one and securing the next one, etc). Opendoor promises to vastly simplify this part, which I think is good.

3) I believe that some of these assumptions are specific of the current market situation, where in most "hot" real estate areas (SF, LA, Seattle, NY, Honolulu) prices are wild, the appetite for real estate has never been stronger, and at the same time there's still many opportunities to make a quick buck by flipping, accessing certain information earlier, etc.

I am not completely sure what will happen when/if the market corrects (might also depend on the size of the correction).

4) In the US the whole process of buying homes, financing them, and potentially use LLCs/Trusts/etc to handle various real estate properties as investment is still very fragmented, and still very dependent on which State you live in and in which State the property resides. There are essentially thousands of different combinations and configurations, and the optimal answer to each can also vary over the years, following updates to the legal code or to tax rules.

All in all, I think that Zillow is trying to carve out another piece of the market, and not necessarily because their current market is showing signs of slowing down (even if that has discussed just recently [0]). Heck, it might even be that Opendoor's outstanding fundraising results might have triggered the decision to go against them.

5) Long term, I think most real estate assets will become fully digital, and handled exactly like a small piece of software - that is, for everything that pertains property and ownership, taxation, transactions, etc. Of course, fixing a leaking pipe will still require a physical intervention :)

For background: I'm co-founder/CEO of a startup which will provide a software platform to digitize real estate assets, and legally transact on properties using APIs. We leverage the Ethereum Blockchain as a global land registry where these transactions are recorded. (I don't intend to use this comment to publicize it, hence no link).

[0]: https://mailchi.mp/12c664b1b53c/realtorcom

Zillow is not to be trusted they have power to manipulate the market prices with their zestimate. Not only that but they're ridiculous leads pricing they charge the agents ($250 for every time you fill out the contact me).

I'm not sure Zestimate actually manipulates the market. Banks don't care about Zestimates when they're handing out loans, they care about appraised value. Zestimates don't recoup the losses on a foreclosure.

Anyone using the Zestimate as a valuation tool is an idiot. It’s helpful for just comparisons, but frequently wrong when it comes to actual transactions.

Next week: Yelp to open chain of restaurant, citing vast knowledge of what makes for a satisfying dining experience. Spokespeople ensure their users that the impartiality of their algorithms will continue. This story brought to you from Google News Reporting.

Definitely not another housing bubble. No way.

The real question is who will be holding the bag after midterm elections.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact