Procore shifts most Analytics 2.0 data to hourly refreshes

Procore rebuilt Analytics 2.0 around OpenSharing, shifting most customer data from a daily API-based extraction cycle to hourly refreshes, Databricks says. Procore’s refresh documentation qualifies that cadence: user activity, adoption and risk tables are among those refreshed daily, while most weekend refreshes occur about every two hours.

According to Databricks, Analytics 2.0 runs a multi-region medallion pipeline. Raw application and integration events land unchanged in a bronze layer, undergo cleaning and validation in silver, and are aggregated and modeled for analytics in gold. Procore consolidates the gold tables in one primary analytics region for distribution. That layer also feeds Procore’s in-app reports and customer extracts, aligning their schemas and naming conventions.

The previous version extracted data from Procore’s public APIs into an Azure-hosted SQL Server backend, Databricks says. Standard refreshes ran daily, and customers paying for additional runs were capped at three refreshes per day. Procore’s Analytics 2.0 overview lists an hourly cadence and says an incremental-refresh feature introduced in April 2025 updates only changed data.

Customers have two delivery paths. Those without Databricks generate an OpenSharing token in Procore and connect through the open protocol. Databricks customers provide a workspace identifier for native sharing between Databricks environments. Databricks says a scheduled job reconciles Procore permissions with sharing entitlements, enforcing access by user and project without manual provisioning.

Procore can scope shares to schemas, views, tables and partitions. With Change Data Feed enabled, recipients can take a full dataset or only records changed since their previous synchronization. Databricks says the system distributes hundreds of curated tables through hundreds of thousands of customer-partitioned shares. It also says Procore expanded the extractable catalog to hundreds of tables in less than half the time needed to build the Analytics 1.0 baseline, effectively doubling the data surface. The opened sources do not independently audit those scale and development-time claims.

Procore product manager Matthew Guptill said OpenSharing increased data availability and visibility and is helping construction teams deliver projects on time, within budget and safely. The statement about business impact is a company claim, not an independently measured result. Databricks also says Procore established four breaking-change releases per year with six weeks’ notice and 12 non-breaking releases per year with three weeks’ notice.