StackOS Community Q&A • best-practice
How to choose between Snowflake, BigQuery, and Redshift for analytics?
We're building a data warehouse for our e-commerce platform (10M+ events/day). Our analytics team of 5 needs to run complex SQL queries, build dashboards, and serve data to ML models. Current stack: Fivetran for ETL, dbt for transformations, Looker for dashboards. Which warehouse would you recommend for this scale and use case?
Top Community Answer
At 10M events/day, all three will handle the volume. The decision comes down to pricing model and ecosystem: **Snowflake** - Best for: Multi-cloud or cloud-agnostic strategy - Pricing: Pay per second of compute (can be unpredictable) - Standout: Zero-copy cloning, excellent data sharing, Snowpark for Python/Java UDFs - With dbt: First-class integration, materialized views supported **BigQuery** - Best for: Google Cloud shops or cost predictability needs - Pricing: Flat-rate slots OR on-demand ($5/TB scanned) - Standout: ML integration (BQML), geospatial functions, free 1TB/month - With dbt: Works well but materialized views have limitations **Redshift** - Best for: AWS-native teams, existing Redshift spectrum users - Pricing: RA3 nodes with managed storage (predictable) - Standout: Redshift Spectrum for S3 queries, AQUA acceleration - With dbt: Solid integration, good for large table scans **Our recommendation:** BigQuery with on-demand pricing. At 10M events/day (~3-5TB/month after compression), you'll likely stay near the free tier for queries. Fivetran + dbt + BigQuery is a battle-tested combination.