Shovels, ReZone, and the Local Data Problem
First, let's get the news out of the way: ReZone, a company Thesis Driven co-founded and incubated, was just acquired by Shovels, a leading provider of permit data and built world intelligence.
The acquisition is a win for Thesis Driven, Shovels, and the ReZone team including Daniel Heller, the company's founder and CEO.
If you'd like to read more about the acquisition, you can check out the press release here.
But today I'd like to talk a bit about why we care so much about this problem, why this acquisition matters, and the relationship between theory and praxis here at Thesis Driven.
A little over a year ago, we ran a thought experiment: what would it take to build a real estate development firm with zero humans? By stringing together automated and AI-powered tools, could computers do a good enough job of site selection, planning, permitting, and construction to deliver a project – and perhaps even a good financial return?
While there are technological gaps today that would make a true "zero-human" scenario effectively impossible, large chunks of the development process – particularly the earliest stages of site sourcing and evaluation – can be meaningfully automated and improved with technology out in the market today. Tools like Deepblocks and Pillar enable high-throughput site analysis while Testfit and Algoma push deeper into massing and design, translating zoning into project details like square footage and unit count.
But as we explored in a letter last year on AI's role in feasibility analysis, what can be built on a given site is often much more complex than the in-place zoning code:
“In many places, it’s not just about the zoning code telling you what size massing blocks you can extrude onto the site,” says Vansice. “There are layers and layers of additional constraints and allowances like density bonuses, affordability requirements, height limiting compatibility, transportation and environmental considerations, and on and on—it’s really complicated to understand what entitlements, yield, and costs are going to look like post-permitting when you’re initially underwriting a deal.”
Worse, many markets heavily rely on discretionary approvals for pretty much every site; the in-place zoning matters far less than what the local council member is likely to support. Without an understanding of what is likely to gain discretionary approval on any given site, any automated site analysis will be incomplete, often woefully so.
Getting that intelligence requires digging deep into the weeds of what actually gets approved, data that can only be gathered in the weeds of planning committee and city council meetings. Which is exactly what ReZone did and what Shovels will now be doing, bridging one of the last true data gaps of automated development.
To be clear, this is hardly a solved problem. Obtaining and structuring the data on zoning approvals and permits is necessary but not sufficient; using it to identify high-potential sites and recommend designs likely to win approval is the holy grail.
Needless to say, I'm excited to see what Daniel, Ryan (Buckley, Shovels' CEO) and the team make happen, and I'll be serving in an advisory role there post-acquisition.
These venture studio shenanigans may seem odd coming from a media company, but they shouldn't surprise longtime Thesis Driven readers. This was never a purely journalistic project; we've built a number of products in our three years that address specific pain points we saw in the market: understanding the real estate industry (our courses and workshops), identifying and selling into real estate operators (GP database), and reaching real estate investors (CapitalStack). While ReZone was our first minority-stake incubation, it won't be our last.
-Brad Hargreaves
