How Curri Cut Data Sync Costs by 50% and Achieved Real-Time Analytics with Estuary
Key Metrics Snapshot
- 50%+ Cost reduction vs. Fivetran
- 12 hours → Instant Stripe payment data sync time
- TB+ Production data moved
"The dream of streaming made simple is how I put it. You give it a go, you're surprised at how easy the configuration was, and then you see it all just work."
— Evan Dixon, Data Engineer at Curri
About Curri
Location
California, United States
Industry
Logistics & Construction Supply Delivery
Goals
Curri aimed to cut unpredictable data integration costs and replace slow, batch syncs with scalable real-time pipelines across 16+ sources to support finance, analytics, and machine learning.
Curri, headquartered in California, is a logistics platform known as the “Uber for construction supplies.” It helps suppliers and contractors deliver building materials nationwide, powered by data-driven operations that support finance, analytics, and real-time decision-making.
Challenge
Curri, the "Uber for construction supplies," faced a data infrastructure challenge that threatened their ability to serve B2B customers effectively. With a small but ambitious data team managing hundreds of production tables and terabytes of data, their analytics setup built on PostgreSQL was creating dangerous dependencies between operational and analytics workloads.
Their initial attempt to modernize involved popular off-the-shelf replication tools like Fivetran and Snowflake's native replication. However, both solutions quickly revealed limitations. Fivetran’s pricing model, based on Monthly Active Rows, introduced unpredictable costs that scaled poorly with Curri's growing operational data. Meanwhile, Snowflake's native replication required extensive manual configuration, could only handle 15 tables at a time, and projected high annual costs just for syncing data — even before factoring in storage and compute fees.
Ultimately, these solutions fell short on both cost-efficiency and scalability, prompting Curri to seek a more modern, real-time approach.
Curri needed a solution that could:
- Sync 16+ data sources in real-time, especially critical payment data from Stripe
- Handle their full production database without manual batching
- Deliver data fast enough for invoicing deadlines
- Reduce operational costs while improving performance
Solution
Estuary Flow transformed Curri's data infrastructure with a genuinely simple approach to streaming data integration. Unlike the complex Kafka-based alternatives that required extensive setup, Estuary's intuitive decoupled "sources → collections → destinations" model allowed Curri to connect their entire production environment during a free trial, without hitting any credit limits.
The implementation was straightforward. Curri's data team connected all 16 data sources, including their 400+ table PostgreSQL database, Stripe, and HubSpot, creating a unified real-time data pipeline to Snowflake. Estuary's flexible sync scheduling enabled smart cost optimization: fast syncs during critical Friday morning invoicing periods and slower syncs during off-peak times.
The solution seamlessly integrated with Curri's existing stack, feeding real-time data into Snowflake for analytics while supporting their Python microservices and machine learning models that calculate fair driver payments based on five years of delivery data.
Results
Estuary delivered transformative results across every metric that mattered to Curri's business. The company achieved a 50% reduction in data integration costs for their entire PostgreSQL replication. More importantly, Stripe payment data that previously took 12 hours to sync now updates instantly, eliminating delays for the finance team during critical Friday invoicing runs.
The performance improvements extended across all data sources, with HubSpot syncs dropping from 3.5 hours to real-time. This newfound speed enabled Curri to build sophisticated real-time applications, including AI models that train frequently on transformed data to calculate optimal driver payments.
"Estuary and Snowflake are like best friends. They speak very well together."
Looking forward, Curri continues to expand their use of Estuary's platform, exploring in-flight transformations and building new real-time analytics capabilities that would have been impossible with their previous batch-based approach. The simplicity and reliability of the solution have transformed how their small data team thinks about streaming: from a complex engineering challenge to an accessible tool for business innovation.