Key Takeaway: Every proptech company starts with a workflow problem and discovers the workflow is commoditizable but the data isn't. I watched it happen from inside Staircase and watched it happen to every competitor — the sequence is predictable to the quarter. The workflow is the acquisition channel. The data is the product.
Call it the Proptech Pivot. Every proptech company starts with a workflow problem. Every one discovers the workflow is commoditizable but the data isn't. Every one pivots to intelligence. The sequence is predictable down to the quarter. I watched it happen from inside, I watched it happen to competitors, and I've watched it happen in every company I've worked with since. The pattern is so reliable that when a proptech founder tells me their competitive advantage is their workflow, I already know what they'll be selling in eighteen months.
Build a tool that digitizes a real estate process. Customers adopt. Competitors show up with similar tools -- because the underlying process is standardized and any competent engineering team can build it. Features converge. Price competition starts. Eighteen months in, you're selling a commodity. Then you look at what you actually have: a unique dataset built from thousands of transactions. The workflow was never your product. It was your acquisition channel for data.
I watched this happen from inside Staircase. I built workflow automation for county data normalization. The workflow was replicable -- any team with enough engineers could scrape county sites and normalize schemas. What wasn't replicable was three years of accumulated data covering 3,100+ counties with quality signals, temporal depth, and entity resolution that only emerge from processing millions of records. The workflow was commoditizable. The county data wasn't.
The evidence is everywhere
CoStar started by sending researchers to physically visit commercial properties. The workflow -- drive to building, measure it, take photos -- was replicable. The database was not. Thirty years of data covering 6 million+ properties. Revenue: $2.7 billion annually. Market cap: $30 billion.
Zillow started as a data play (the Zestimate), tried to become a workflow company through iBuying, lost $881 million in Q3 2021 alone, and retreated to data. Margins went back to 70%+. The market told them what they were.
Compass built an agent-facing workflow platform and spent billions acquiring brokerages. The tools are fine but not defensible. $6 billion in revenue, negligible profit. Workflow without proprietary data is a low-margin business.
Why the pattern holds
Real estate workflows are complex but finite. A closing involves defined steps, defined participants, defined documents. The barriers to good workflow tools are execution and go-to-market -- advantages that erode. The barriers to good datasets are time, transaction volume, and data quality -- advantages that compound.
Three properties make real estate data uniquely valuable. Fragmentation -- 3,600+ counties with no standardization means aggregation itself creates value. Opacity -- most transactions are private, so any company accumulating transaction-level data has information the broader market doesn't. Temporal depth -- real estate cycles run 7-10 years, so a dataset covering three cycles is predictive. You can't shortcut time.
The law
In real estate technology, the workflow is the acquisition channel and the data is the product. Any proptech workflow company will either pivot to data, get acquired for its dataset, or compete on price until margins go to zero. Those are the three exits.
Instead of spending your engineering budget on UI polish, invest in data architecture. The schema you choose, the metadata you capture, the historical data you retain -- these determine whether you have a commodity SaaS product or a data asset that compounds. When you're making build-vs-buy decisions, ask one question: does the workflow generate proprietary data? If yes, build. If no, buy.