QueryFuser Pricing & BigQuery Savings
QueryFuser is free during beta. After beta, pricing will be tied directly to the value we deliver.
- Unlimited BigQuery projects
- Unlimited users
- All merge features — query fusion, fallback, table overlap
- Per-project merge configuration
- Max merge group size controls
- Full merge & query log retention
- Admin dashboard
Usage-Based Pricing
When QueryFuser moves out of beta, pricing will be tied directly to the value we deliver. We’re evaluating two models:
Per Query Merge
Pay a small fee for each successful merge group. If QueryFuser merges 10 queries into 1, that counts as one merge event. You only pay when merging actually happens — pass-through queries are always free.
Based on Bytes Saved
Pay based on the actual bytes you save. QueryFuser already logs bytes_processed for every merged query, so we can calculate your exact savings compared to individual execution.
Both models ensure you only pay when QueryFuser delivers real value. During beta, all features are free with no commitment.
How Much Could You Save?
Savings depend on three factors: how many tiles a dashboard fires at the same tables, how much their referenced columns overlap, and the size of those columns. BigQuery is columnar — it scans only the columns each query references, not the full table.
| Scenario | Tiles | Solo Cost | Merged Cost | Savings |
|---|---|---|---|---|
| ① Star schema, INT64 keys Fact table + 3 dims (100–1,000× smaller) |
8 | 16 units | 4 units | 75% |
| ② Star schema, STRING keys Same, but keys are 4× size of INT |
8 | 40 units | 13 units | 68% |
| ③ Single table, 5 string dims 100 M rows, no joins, each dim = 4× amount |
8 | 40 units | 21 units | 48% |
Baseline: 8-tile dashboard, 100 M-row fact table, 3 dim tables. 1 unit = one INT64/FLOAT64 column (800 MB). STRING columns at ~32 bytes = 4 units. Dashboards with more tiles or fewer dims save even more. Percentages stay valid regardless of table size, BigQuery pricing tier, or caching.
Key Factors
Three things determine how much QueryFuser saves you. Understanding them helps set realistic expectations.
Concurrency
More tiles hitting the same tables = higher savings. A typical 8-tile dashboard with INT keys saves 75%. Dashboards with 12+ tiles can exceed 85%. When multiple users load the same dashboard, savings multiply further.
Column Overlap
Star schema with 3 join keys: merged query needs 4 columns (75%). Denormalized table with 5 string dimensions: merged query needs 21 units (48%). Fewer distinct columns in the merge = higher savings.
Partitioning
Date-partitioned tables benefit when tiles share the same partitions. QueryFuser deduplicates overlapping partition reads. Non-overlapping filters already prune optimally via partitions, so savings depend on how much partition overlap exists.
Typical 8-tile dashboard: star schema with INT64 keys saves 75%, STRING keys 68%, denormalized table with 5 string dims 48%. More tiles and fewer distinct columns push savings higher.
Common Questions
Does QueryFuser add to my BigQuery bill?
No. QueryFuser only reduces the number of queries sent to BigQuery. It does not add any queries, intermediate tables, or materialized views.
What happens if a merged query fails?
QueryFuser falls back to executing each query individually. Queries that would have succeeded on their own still succeed. Queries that would have failed still return their original error. No data is lost.
How long will the beta last?
We don't have a fixed end date. We'll announce pricing well in advance before any charges begin. All beta users will get early-adopter benefits.
Can I use flat-rate or editions pricing with QueryFuser?
Yes. On flat-rate pricing you pay for slots, not bytes scanned. QueryFuser still helps by reducing the number of concurrent slot reservations needed, freeing capacity for other workloads.
Do I need to change my SQL or application code?
No. QueryFuser is transparent. Point your BI tool at QueryFuser's host and port instead of BigQuery directly. Everything else stays the same.