How to reduce BigQuery costs for a data warehouse with high query volume?

All answers to this question.

The fastest way to save money in BigQuery is to implement partitioning and clustering. By partitioning on a date column, your queries will only scan the specific days required rather than the entire table. Clustering further organizes the data within those partitions, which can reduce the "bytes scanned" significantly. Also, ban the use of SELECT * across the team—analysts should only select the columns they need. Finally, check your "Long-term Storage" status. Data that hasn't been modified in 90 days automatically drops in price by 50%, so try to avoid re-writing your entire history unnecessarily.

   Answered 2025-08-04 by Kimberly Adams

Have you explored the "BigQuery BI Engine" to cache frequently accessed data, or are your users constantly querying the raw tables for every single dashboard refresh?

   Answered 2025-08-16 by Jeffrey Lewis

• Jeffrey, we haven't tried BI Engine yet! I just read that it’s an in-memory analysis service that can significantly speed up dashboard tools like Looker. If it reduces the number of times we hit the actual storage engine for the same repetitive queries, that could be a huge cost saver. I’m also going to set up "Query Quotas" at the user level today. It’s a bit of a "tough love" approach, but it will force our analysts to be more mindful about the efficiency of the SQL code they are writing.

     Commented 2025-08-18 by Anthony Wright

Switching to "Editions" (Standard or Enterprise) can help if you want predictable costs. You pay for "compute slots" rather than the number of bytes you scan.

   Answered 2025-08-20 by Susan Baker

• I second that. If you have high query volume but predictable schedules, slot-based pricing is almost always cheaper than the "on-demand" model for large enterprises.

     Commented 2025-08-22 by Patricia King