Shovels prioritizes control over volume
Shovels
This drop shows that the hard part of Shovels is not storing permit records, it is continuously collecting and normalizing messy local government data at national scale. Shovels still covered more than 178 million permits across over 2,770 jurisdictions and roughly 85% of the U.S. population after the transition, but the lower count revealed how much historical breadth had depended on outside licensed feeds rather than only its own collectors.
-
Shovels chose control over headline volume. By moving to fully first party collection in 2025 and 2026, it removed upstream licensing fees and gained direct control over refresh timing, schema mapping, and quality assurance, but it also took full responsibility for every broken portal, missing archive, and format change.
-
Permit coverage is unusually fragile because there is no national system. More than 20,000 U.S. jurisdictions publish records through different portals, field names, taxonomies, and update schedules, so a lower record count can reflect stricter deduping, cleaner normalization, or loss of hard to access historical backfill, not just weaker current collection.
-
This is also where Shovels differs from broader incumbents like ATTOM and Cotality, which bundle permits inside larger property data products, and from BuildZoom, which says it has more than 326 million permits and adds marketplace data on contractors. Shovels is betting that a neutral API and warehouse product matters more than the biggest raw archive.
Going forward, the prize is a dataset that is smaller on paper but more dependable in production. If Shovels keeps expanding first party coverage while linking permits with contractors, properties, and pre permit decisions, it can become the system enterprises use for live construction intelligence rather than a static archive of old municipal records.