Procore Rebuilds Analytics Product on Databricks OpenSharing
Procore says it replaced its API-based analytics tool with Databricks OpenSharing, removing a daily refresh limit and doubling its catalog of extractable tables.
Respuesta rápida
How did Procore use Databricks OpenSharing to rebuild its construction analytics product?
Procore says it rebuilt Procore Analytics on Databricks OpenSharing, replacing an API-driven system capped at three daily refreshes. The company reports the new version distributes hundreds of curated tables through per-customer partitioned access, lifts the refresh ceiling to hourly, and doubled the catalog of data customers can extract.
Claves
- Procore says it replaced an API-driven analytics backend, capped at a maximum of three refreshes per day, with Databricks OpenSharing.
- According to Procore, the new system removes the daily refresh ceiling, letting customers pull data on an hourly cadence.
- Procore reports OpenSharing distributes hundreds of curated tables while maintaining partitioned, per-customer access across hundreds of thousands of distinct data shares.
- The company says the rebuilt product, Procore Analytics 2.0, expanded its table catalog to hundreds of tables in less than half the time it took to build the original catalog, doubling the surface area of extractable data.
- Procore says the architecture uses Databricks' Unity Catalog for governance and supports both Databricks-to-Databricks sharing and an open protocol for customers on other platforms.
Procore moves analytics product to Databricks OpenSharing
Procore, a construction management software provider, says it has rebuilt its customer-facing analytics product, Procore Analytics, on top of Databricks' OpenSharing data-sharing technology, replacing an earlier version that relied on public API extractions into a SQL Server backend.
According to Procore, the original version of the product — which the company calls Procore Analytics 1.0 — was limited to a single daily data refresh, with a maximum of three refreshes per day even for customers who paid for additional updates. Procore says that architecture also created cost and reliability problems at scale, since pulling data for tens of thousands of customers through public APIs required deep recursive calls that were prone to timeouts and rate limits.
What changed in the new version
Procore says the new version, Procore Analytics 2.0, is built on a medallion architecture inside Databricks, with raw data processed through bronze, silver, and gold layers before being distributed to customers through OpenSharing. The company says bronze, silver, and gold processing runs across multiple regions, with gold tables consolidated into a single primary analytics region for distribution.
Procore describes two ways customers receive data under the new system. Customers without a Databricks workspace generate an access token inside the Procore application and connect using the open OpenSharing protocol, while customers who already run Databricks receive shares natively through Databricks-to-Databricks sharing. The company says a scheduled job automatically reconciles each user's permissions from the Procore application into OpenSharing entitlements, enforcing access per user and per project without manual provisioning.
Procore says data governance for the new system runs through Databricks' Unity Catalog, and that tables have Change Data Feed enabled so recipients can sync full datasets or only the data that has changed since their last pull. The company also says it has standardized on a release cadence of four breaking-change releases per year, each with six weeks of advance notice, plus twelve non-breaking releases per year with three weeks of notice.
Reported results
Procore says the shift to OpenSharing lifted the daily refresh ceiling, allowing customers to pull data on an hourly basis or on whatever cadence their downstream tools support. The company reports that while the original product took several years to build out its baseline catalog of available tables, the new version expanded that catalog to hundreds of tables in less than half the time, which Procore says effectively doubled the amount of data customers can extract at scale.
Procore also says the new architecture consolidated what had previously been two parallel data pipelines — one for customer-facing analytics and one for in-app reporting — into a single pipeline that feeds both. The company says this means a product team can ship a new feature and have its associated data available for customer extraction almost immediately, rather than waiting for a separate analytics pipeline to catch up.
Source: Databricks Blog, "How Procore is powering construction analytics with OpenSharing," published October 29, 2026.
Preguntas frecuentes
- What problem was Procore trying to solve with the new analytics architecture?
- Procore says its original analytics product, built on a SQL Server backend fed by API extractions, was limited to a maximum of three refreshes per day and suffered from cost and reliability issues when pulling data for tens of thousands of customers through public APIs.
- What is OpenSharing and how does Procore use it?
- According to the announcement, OpenSharing is an open data-sharing protocol on Databricks that Procore uses to distribute hundreds of curated tables directly to customers, maintaining per-customer partitioned access and supporting both Databricks-native and non-Databricks recipients.
- What measurable results does Procore report from the change?
- Procore says the move doubled the catalog of extractable tables compared to the original product and removed the once-a-day refresh limit, letting customers pull data multiple times per day.
- How does Procore handle data governance and permissions under the new system?
- Procore says sharing runs through Databricks' Unity Catalog, with a scheduled job that reconciles each user's permissions from the Procore application to enforce per-user, per-project access automatically.