I Was Wrong About Zillow

Which Is Exactly Why You Should Be Paying Attention

In 2024, I wrote a piece arguing that Zillow was evolving from a real estate search portal into something far more powerful: a vertically integrated machine capable of reshaping the entire housing industry.

A few parts of that article aged well.

A few parts aged like gas station sushi left on the dashboard of a 2007 Escalade in August.

So let’s clean this up.

Zillow did not successfully conquer America through iBuying. In fact, Zillow Offers became one of the most spectacular corporate faceplant in modern real estate history.

Turns out houses are not stocks. Humans are inconvenient. Foundations crack. Sewer lines collapse. Appraisers exist. Roofs leak. People panic-buy during low interest rates and panic-sell during high ones.

Algorithms hate unpredictability.

Unfortunately for humanity, Silicon Valley sees unpredictability not as a warning sign, but as a software challenge.

So while I was wrong about how Zillow would attempt to dominate real estate, I may have underestimated something much larger:

The portal war was never really about buying houses.

It was about owning attention, data, consumer behavior, and eventually the transaction itself.

And that story is still very much alive.

Zillow Is Not Becoming the Broker

It Is Becoming the Operating System

This is the important distinction.

The original fear among agents was that Zillow wanted to replace REALTORS entirely. That turned out to be overly simplistic.

What Zillow appears to want is something far more profitable and far less labor-intensive:

To become the infrastructure layer sitting between consumers and everyone else.

Consumers search on Zillow.
Agents advertise on Zillow.
Showing appointments increasingly flow through Zillow-owned systems.
Lead generation flows through Zillow.
Mortgage referrals flow through Zillow.
Rental data flows through Zillow.
Behavioral data flows through Zillow.

Soon enough, transaction management, AI recommendations, financing prompts, predictive analytics, insurance offers, renovation estimates, and automated negotiations all flow through Zillow too.

You do not have to replace the agent if you can tax the entire ecosystem.

That is not a brokerage model.

That is an extraction model.

The Quiet Death of the Independent Website

Remember when brokerages believed their websites were their competitive advantage?

That was adorable.

Most brokerage websites today function like digital pamphlets floating in the shadow of giant portals that spend billions optimizing search, AI, consumer retention, and behavioral tracking.

Consumers no longer simply search for homes.

They enter ecosystems.

And ecosystems are sticky.

Spotify does not just play music.
Amazon does not just sell products.
Apple does not just make phones.

And Zillow no longer just displays listings.

It shapes consumer perception itself.

The company controlling the interface often becomes more powerful than the companies controlling the inventory.

Ask hotels how that worked out with Expedia.

Ask restaurants how they feel about DoorDash.

Ask drivers how much they love Uber taking a percentage of every ride while insisting they are empowering local transportation entrepreneurs.

Nothing says freedom like paying tribute to a software company before you can earn a living.

Artificial Intelligence Changes the Equation

This is where things become genuinely interesting.

And slightly dystopian.

The next generation of AI will not merely answer questions like:

How many bedrooms?

It will answer questions like:

Which homes fit my psychological profile, spending behavior, commuting tolerance, political preferences, social status aspirations, health concerns, school priorities, and estimated future income trajectory?

That sounds insane until you realize most tech companies already collect enough behavioral data to attempt it.

AI does not simply organize information.

It predicts behavior.

And predictive systems become extraordinarily powerful when paired with weak competition and massive datasets.

Now imagine a future where:

AI recommends neighborhoods.
AI recommends financing.
AI recommends insurance.
AI recommends contractors.
AI recommends pricing strategy.
AI negotiates transaction terms.
AI predicts which homeowners are most likely to sell before they even know it themselves.

At that point, real estate stops looking like a marketplace.

It starts looking like a managed behavioral funnel.

Weak Antitrust Enforcement Created This Environment

For decades, America basically adopted the antitrust philosophy of:

If consumers are not visibly paying more today, who cares?

That framework worked wonderfully if your goal was creating trillion-dollar technology companies with enough market influence to rival small governments.

Modern monopolies do not always raise prices immediately.

Sometimes they centralize data.

Sometimes they control distribution.

Sometimes they quietly become unavoidable infrastructure.

And by the time regulators notice, half the economy is renting oxygen from four apps and a cloud server farm in Northern Virginia.

Real estate may be heading down a similar path.

Not because Zillow is uniquely evil.

Because the economic incentives pushing toward consolidation are overwhelming.

Data advantages compound.
AI advantages compound.
Consumer attention compounds.
Network effects compound.

And local expertise does not scale nearly as efficiently as software.

The Irony Nobody Wants to Admit

The more powerful AI becomes, the more valuable true local expertise may actually become at the highest end of the market.

Especially in places like Southwest Michigan.

Lakefront property is not standardized.

Resort markets are not rational.

Luxury buyers are not spreadsheets.

Algorithms struggle with nuance.

A bluff-top estate with erosion concerns, private beach rights, EGLE restrictions, architectural significance, generational family history, and a Chicago buyer seeking a psychological escape from urban life is not the same as pricing a three-bedroom subdivision house in Phoenix.

And wealthy buyers know it.

At the high end, people are not merely purchasing square footage.

They are purchasing judgment.

Discretion.

Access.

Interpretation.

Protection from mistakes.

That becomes more valuable, not less, in a world drowning in automated information.

The Future Probably Looks Hybrid

AI will absolutely replace portions of real estate.

A lot of portions.

Lead routing.
Search.
Scheduling.
Pricing analysis.
Document review.
Marketing copy.
Customer follow-up.
Ad targeting.

Much of the industry still operates like it is 2009 with better headshots.

That will not survive.

But neither will the fantasy that housing can become fully automated like ordering paper towels on Amazon.

Homes are emotional.

Messy.

Hyper-local.

Financially consequential.

Politically regulated.

Psychologically loaded.

Which means the future probably belongs to people and companies capable of combining technology with actual expertise instead of pretending one eliminates the other.

The danger is not AI itself.

The danger is allowing a handful of companies to become the gatekeepers through which nearly every housing decision flows.

Because once infrastructure becomes centralized, independence becomes performative.

And that is usually when the free market crowd suddenly discovers they have been living inside a company town with better UX design.

Paul Zubrys
Southwest Michigan Real Estate since 1987
Low volume. High touch. Discreet advisory for exceptional buyers.
(269) 783-5253 office, home, mobile
Send me a message, question, or request.