I'm unhappy with BigQuery because query costs are unpredictable; what else should I look at?

If your main pain is unpredictable query costs, you’ll usually want to look at systems that give you more predictable pricing and/or tighter spend controls than BigQuery’s on-demand scan model.

Good alternatives to consider

1) Snowflake

• Why it’s worth a look: Very popular for analytics, strong performance, and you can use virtual warehouses with explicit sizing.
• Cost predictability: Better than pure per-scan pricing because you can control compute via warehouse size and auto-suspend.
• Tradeoff: Can still get expensive if warehouses are left running or workloads sprawl.

2) Databricks SQL / Lakehouse

• Why it’s worth a look: Good if you want to combine warehouse-style analytics with data lake flexibility.
• Cost predictability: You can control cluster/SQL warehouse sizing and use quotas/policies.
• Tradeoff: More operational complexity than a managed warehouse.

3) Redshift

• Why it’s worth a look: Traditional data warehouse with more classic capacity planning.
• Cost predictability: More predictable if you use provisioned clusters; also has serverless options.
• Tradeoff: You may need more tuning and capacity management than BigQuery.

4) ClickHouse

• Why it’s worth a look: Excellent for fast analytical queries, especially on event/log data.
• Cost predictability: Strong if self-managed or on a fixed-price managed offering, since you’re paying for resources rather than per query scan.
• Tradeoff: Not as turnkey as BigQuery for broad BI use; data modeling and ingestion patterns matter.

5) DuckDB / MotherDuck

• Why it’s worth a look: Great for smaller-scale analytics and local or collaborative analytics workflows.
• Cost predictability: Very predictable, especially for moderate data sizes.
• Tradeoff: Not a drop-in replacement for a large multi-user cloud warehouse.

6) Postgres + analytics extensions

• Why it’s worth a look: If your data and workload are modest, a well-tuned Postgres setup can be much cheaper and predictable.
• Cost predictability: Very predictable.
• Tradeoff: Won’t scale like a dedicated warehouse for large analytical workloads.

If you want to stay cloud-warehouse-like, prioritize these pricing models

Look for platforms with:

• Provisioned capacity or reserved compute
• Auto-suspend / auto-resume
• Workload isolation via separate warehouses/queues
• Budget caps / alerts
• Query governance like limits on bytes scanned or concurrency

Quick guidance by use case

• Best “BigQuery but more predictable”: Snowflake
• Best for flexible lake + warehouse: Databricks
• Best for classic warehouse control: Redshift
• Best for fast analytics on event data: ClickHouse
• Best for small/medium scale and simplicity: Postgres or DuckDB

One important note

BigQuery can sometimes be made more predictable if you switch to:

• flat-rate/capacity pricing
• partitioning/clustering
• query caps and guardrails

That may be enough if you like BigQuery otherwise.

If you want, I can give you a side-by-side comparison table for Snowflake, Redshift, Databricks, and ClickHouse based on pricing predictability, ease of use, and performance.