---
title: The Ideas That Never Get Built | Founders Field Notes No. 06
description: "Why so many real estate ideas never get built, or quietly stall: cheap software keeps running into expensive, fragmented MLS data."
image: https://www.pickup-studios.com/hubfs/field-notes-share-card.png
---

[Skip to content](https://www.pickup-studios.com/field-notes/the-ideas-that-never-get-built#main-content)

[![Pickup Studios](https://www.pickup-studios.com/hs-fs/hubfs/Pickup-LogoAsset%201@3x.png?width=178&height=53&name=Pickup-LogoAsset%201@3x.png)Homepage](https://www.pickup-studios.com?hsLang=en)

- [Services](https://www.pickup-studios.com/services?hsLang=en)
- [Blueprint](https://www.pickup-studios.com/innovation-blueprint?hsLang=en)
- [Insights](https://www.pickup-studios.com/blog?hsLang=en)
- [Field Notes](https://www.pickup-studios.com/field-notes?hsLang=en)
- [About](https://www.pickup-studios.com/about-us?hsLang=en)

Get started

- [Services](https://www.pickup-studios.com/services?hsLang=en)
- [Blueprint](https://www.pickup-studios.com/innovation-blueprint?hsLang=en)
- [Insights](https://www.pickup-studios.com/blog?hsLang=en)
- [Field Notes](https://www.pickup-studios.com/field-notes?hsLang=en)
- [About](https://www.pickup-studios.com/about-us?hsLang=en)

Get started

[Founders Field Notes](https://www.pickup-studios.com/field-notes?hsLang=en)

No. 06

Volume I · November 2026Field Note

# The Ideas That Never Get Built

Some never get built. Others get built and quietly stall. In real estate, the reason is often the same: the data underneath them.

**John Fischer**, Founder, Pickup Studios[Download PDF](https://www.pickup-studios.com/hubfs/Pickup-Studios-Field-Notes-No-06.pdf?hsLang=en)

A number of years ago, we worked with a client on a showings prototype with a simple premise. You could text a number to see a house.

No app to download. No portal to log into. A buyer, or a buyer’s agent, would send a text, and that message would go wherever the listing agent wanted it to go: to the agent, or straight to the seller. If the listing agent wasn’t using the service, they still got the text, along with an easy way to start. The idea was that requesting a showing could be as simple as sending a message to a friend.

The prototype worked. The client had access to their local MLS, which held what we needed: listing information, showing instructions and agent contact details. A large regional brokerage was talking with us about testing it with their agents. By most measures, it was the moment a small idea starts to become a real one.

Then we did the math on everything beyond that one market. There are now roughly 500 MLSs in the United States. That number shrinks a little each year as some of them merge, and there were more still when we were working on this. Each has its own agreements, its own rules about how data can be used and its own way of handling showing instructions. To make a text-first showing service work nationally, we would have had to negotiate access, map data and stay in compliance hundreds of times over before a single agent outside the first market could use it.

We stopped. The idea worked and people wanted it, but the data was a moat we couldn’t see a way across. As far as I know, nothing quite like it exists today.

§

## The rental version of the same problem

An earlier project ran into a different version of the same wall. I was a strategic advisor to a startup trying to make it easier to find a rental in New York City. Like many startups, their plan was to start in one place, and then expand.

Rentals are an unusual corner of real estate. In most markets, agents aren’t involved in finding an apartment, and there’s no commission to fund anyone’s time. Partly because of that, there’s no rental equivalent of the MLS: no single, reasonably complete source of what’s available, at what price, right now. A few rental platforms have done well, mostly by building inventory the slow way, one landlord and one property manager at a time. Many others haven’t.

That startup could build a good product. What it couldn’t easily build was the data underneath it.

§

## The expensive part isn’t the software

That’s the lesson I keep coming back to. In this industry, the costly part of a new idea usually isn’t the code. It’s access to the information the code needs.

That gap is getting wider. In [Field Note No. 04](https://www.pickup-studios.com/field-notes/ai-will-flatten-competitive-advantage?hsLang=en), I wrote about how AI is making capabilities cheap. A working prototype now takes a fraction of the time and money it took when we built ours. A small team, or one determined agent, can put something real in front of users in weeks.

So we’re likely to see more prototypes than ever. Whether we see more products depends on what those prototypes can connect to. When data is fragmented, licensed market by market or held privately, cheap software runs straight into expensive data. Some ideas never get past the whiteboard. Others get built, find early users and then quietly stall, because growing past one market costs more than the idea can earn.

> Cheap software runs straight into expensive data.

It’s fair to say the industry has made real progress here. Data standards have made listing information far more consistent than it used to be. But a common format doesn’t solve access, and access is still negotiated one organization at a time.

The industry clearly wants new ideas. There are real estate accelerators, venture arms and conference pitch stages built specifically to find and fund them, and they do valuable work. They can give a small team capital, advice and an audience. What they can’t do is make the data easier to reach. A startup can win a pitch competition and still face hundreds of separate negotiations the next morning. And when fewer teams can get that far, there’s simply less for those programs to find.

§

## The debate is happening between large players

Control of real estate data is one of the most talked-about issues in the industry right now. Compass and Northwest MLS settled a 16-month legal fight in August. Other disputes over private listings, coming-soon rules and who can show what to whom are still working their way through MLSs, portals and courts.

I’m not going to weigh in on who is right in any of them. The people involved have real arguments about seller choice, consumer access and fairness, and reasonable people land in different places.

What I notice is who is in the room. The conversation is mostly among large brokerages, portals and the organizations that run the MLSs. The small builder trying something new usually isn’t part of it, and their stake is easy to miss. Whatever the outcome, every rule about who can access listing data, in what form and at what cost, also decides what a three-person startup can try next year.

§

## More than marketing

There’s another reason to pay attention, and others have written about it more thoroughly than I can here. A shared record of listings and sales does more than help homes get marketed. Appraisers and lenders lean on it for comparable sales. Some have argued that if more activity happens outside that shared record, the information used to value homes gets thinner or less reliable.

I mention it to make a broader point. The MLS functions as shared infrastructure, and infrastructure tends to support uses nobody had in mind when it was built. New products are one of those uses.

§

## Innovation needs a lot of attempts

This is the part I feel most strongly about.

Innovation doesn’t come from a few good ideas. It comes from many ideas, products and prototypes getting built, tested, compared and, most of the time, abandoned. The showing tool that eventually changes how agents work will probably be someone’s fifth or fifteenth attempt, built on lessons from the ones that didn’t make it. Take away the ability to build those early attempts and you don’t only lose them. You lose what would have come after them.

When that happens, the tools that already exist face less pressure to improve. I saw this with showing systems at the time. The established options were clunky, and they had little reason to change, because nobody small enough to be nimble could get far enough to compete with them.

§

## Who else needs to build on it

So when leaders think about data, the question can’t only be who owns it or who controls it. It’s also who else needs to build on it, and what this industry loses if they can’t.

Some of the most useful ideas in real estate will come from people who aren’t large enough to negotiate with 500 organizations. The question is whether they’ll get far enough to find out if they’re right.

### Get the next one

New Field Notes go to subscribers first, usually a week or two before they appear here.

[Subscribe](https://www.pickup-studios.com/field-notes#subscribe)

### Keep a copy

This Field Note is also available as a PDF, for saving or passing along.

[Download PDF](https://www.pickup-studios.com/hubfs/Pickup-Studios-Field-Notes-No-06.pdf?hsLang=en)

[← No. 05 · The Middle Is Getting Harder to Defend](https://www.pickup-studios.com/field-notes/the-middle-is-getting-harder-to-defend?hsLang=en)[All Field Notes →](https://www.pickup-studios.com/field-notes?hsLang=en)

[![Pickup Studios](https://www.pickup-studios.com/hs-fs/hubfs/Secondary%20Blue%20MarkAsset%204@6x.png?width=44&height=44&name=Secondary%20Blue%20MarkAsset%204@6x.png)](https://www.pickup-studios.com/?hsLang=en)

Strategy and innovation for the real estate brokerage ecosystem.

Work

[Services](https://www.pickup-studios.com/services?hsLang=en)[Innovation Blueprint](https://www.pickup-studios.com/innovation-blueprint?hsLang=en)[Contact](https://www.pickup-studios.com/contact-us?hsLang=en)

Thinking

[Insights](https://www.pickup-studios.com/blog?hsLang=en)[Field Notes](https://www.pickup-studios.com/field-notes?hsLang=en)[About](https://www.pickup-studios.com/about-us?hsLang=en)

Connect

[LinkedIn](https://www.linkedin.com/company/pickup-studios)[hello@pickup-studios.com](mailto:hello@pickup-studios.com)

© 2026 Pickup Studios Inc.[Privacy Policy](https://www.pickup-studios.com/privacy-policy?hsLang=en)