BigQuery: bytes scanned is the meter

On demand BigQuery bills by bytes scanned, so a query that reads a whole table when it needs one day of data pays for the whole table. Partitioning by date and clustering by common filter columns lets BigQuery prune to the relevant data, often cutting bytes scanned by an order of magnitude on selective queries.

The single most expensive pattern is the unbounded query run on a schedule. Finding and fixing the queries that scan the most is usually the fastest BigQuery saving available.

On demand versus capacity pricing

BigQuery also offers capacity pricing with reserved or autoscaling slots, billed for compute capacity rather than bytes scanned. Spiky exploratory workloads often stay cheaper on demand, while steady high volume query estates usually cost less on capacity. Editions and autoscaling slots let you match capacity to demand within the capacity model.

Picking the wrong mode is a recurring overpay. Measure your scan profile before committing to either, and revisit it as query volume grows.

Storage tiers and materialized views

BigQuery storage drops to long term pricing automatically for tables not edited for ninety days, and Cloud Storage classes nearline, coldline, and archive price progressively cheaper for data read less often. Cold data left in standard storage overpays every month.

Materialized views precompute and store the result of expensive aggregations, so repeated dashboard and report queries read the small materialized result instead of rescanning the base table. For high frequency aggregations this turns a large recurring scan into a small one.

A worked example

Indicative figures, verified against the client's billing data, anonymized. A Fortune 500 retailer partitioned its largest tables, moved a steady estate to capacity pricing, and added materialized views.

Indicative monthly BigQuery and storage spend before and after
ChoiceBeforeAfter
Unpartitioned tables, full scans28,000 USD9,000 USD
Steady estate on demand19,000 USD12,000 USD on capacity
Cold data in standard storage7,000 USD2,500 USD tiered
Repeated dashboard aggregations11,000 USD3,000 USD materialized

Your next step

Find your heaviest scans first, partition and cluster, then match pricing mode and tier the cold data. For the full method read the GCP cost optimization guide, and for neighbouring detail see BigQuery cost optimization end to end and the economics of GCP data lakes. To apply it, our GCP cost optimization service turns the review into verified savings, and you can request a free trial.

Independent · buyer-side

Put a defensible number on your cloud spend.

No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.

Buyer-side intelligence, monthly.

The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.