Since everything in real estate is mystified by people who profit from the mystification, you'd be hard-pressed to understand why the MLS is such a mess. Ask around and you'll get a different answer from everyone. Regulatory capture. Too many standards. Legacy tech systems. etc

I've spent the last year consulting for a real estate tech startup, which means I've spent the last year navigating around MLS data. Looking at API docs from MLS aggregators, reading tri-party agreements, learning what the acronym IDX means and why it makes me want to throw up. After a while, I got curious about how exactly we ended up with this clusterfuck.

The MLS (Multiple Listing Service) is fragmented at two levels: technical and governance. And when I say MLS, I mean both American and Canadian systems. The two countries have similar real estate market dynamics and the same basic organizational mess. Technically, you've got 500+ local systems with different schemas, field definitions, and data formats. "Pending" means something different in Phoenix than it does in Philadelphia.

The technical layer is getting fixed. RESO is a real standard and adoption is growing. But the governance layer is far worse. Each of those 500+ MLSs is a separate organization requiring its own licensing agreement to access their data. Different terms, different fees, different timelines. You can have perfect APIs and still spend months in contract purgatory.

But how did we get here?

Why MLSs Exist

The origin is in San Diego, 1885 when brokers started sharing property lists so any member could bring a buyer and get paid. One office rarely had enough buyers for its own listings, and one buyer rarely found the right property in a single office's inventory. Broker cooperation made everyone money.

The idea caught on. Local realtor boards across the country adopted the same basic model over the following decades. Meetings, bulletin boards, printed sheets, eventually weekly books with photos. The MLS was a coordination mechanism: you share your inventory, I share mine, we split the commission when a deal closes. Everyone wins.

MLSs were centralized within a local market, but they were never designed for cross-market interoperability. Their jurisdiction matched the politics of local realtor associations. This made perfect sense in 1885 when nobody was buying a house in Phoenix from Philadelphia. It makes almost no sense now, but the organizational structure persists because the people benefiting from fragmentation would have to vote against their own interests to change it.

The Tech Is (Slowly) Getting Fixed

For most of the 20th century, the MLS ran on paperwork. Every piece of tech layered on top just reinforced the parochial rules already in place. Weird codes for property features, unyielding data fields. When those same systems went digital, they carried over the insularity and fragmentation.

Realtor.com launched in 1995. Listings became a consumer product instead of a members-only resource. Once buyers could search from home, "hiding listings" started feeling illegitimate. IDX (Internet Data Exchange) formalized broker reciprocity online—brokers could display other participants' listings on their own sites.

But it was largely the same structure as we had with paper. You share MY listings, I share YOURS, and we split the commission when a deal closes.

And with the work from RESO, we're finally getting a larger number of MLSs to adopt a common standard data schema. The tech im sure will eventually catch up, so you no longer have to deal with field normalization headaches.

The Fragmentation Is Sticky

Each local MLS is a little fiefdom. Board members, staff, dues revenue, political power within the local realtor association. Consolidating into a national licensing framework would mean those people voting themselves out of jobs and relevance. That's not going to happen voluntarily.

There has been some consolidation. There are 50% fewer MLS organizations than there was just 10 years ago. But "50% fewer" still means 550. The consolidation that's happened is the easy kind—merging adjacent markets, reducing obvious redundancy. The hard kind—a single national framework that lets anyone access data under standard terms—would require the turkeys to vote for Thanksgiving.

And there's no one to force it. In the US, NAR, or in Canada, CREA, could theoretically mandate universal data access agreement, but both are controlled by the same people who run the local boards. The people who would benefit from a single agreement—tech companies, startups, consumers—have no seat at the table. The people who benefit from fragmentation control the table.

The free market doesn't work here because you can't build a competing MLS to profit from the dysfunction. The barriers to entry aren't technical—they're mafia-esque. You need broker participation, and brokers are required to use the local MLS by their associations.

The tech is getting fixed. But absent regulatory compulsion or a structural shift in who captures value, the governance layer's fragmentation will persist.