Cut your BigQuery and Databricks bill without slowing anyone down
We read 30 days of query, job and cluster history, find the queries, tables and clusters that cost the most, and show your data team exactly how to fix them. Fixed price, results in about two weeks.
Your team's time: about 2 hours in total, for setup and the readout.
- Price
- From $2,500, fixed
- Timeline
- About 2 weeks
- Looks at
- 30 days of history
- Reads
- Metadata, never table data
Is this right for you?
A good fit if
- You spend $5,000 or more a month on BigQuery or Databricks and it is growing faster than your data team.
- You are deciding between BigQuery on-demand and editions, or whether to buy slot commitments.
- A handful of dashboards, dbt models or notebooks seem to drive most of the bill, and nobody is sure which.
- Your data team wants to keep shipping, without a finance-led freeze.
Probably not for you if
- You run mainly on Snowflake or Redshift. We do not audit Snowflake today.
- Your Databricks workspaces do not have Unity Catalog system tables enabled yet.
- You want someone to change reservations or clusters without your team reviewing it.
What you walk away with
The real cost drivers, named
Cost by project, user, service account and query pattern, or by workspace, cluster, job and warehouse, so you know exactly where the money goes.
The right pricing model
A clear answer on on-demand versus editions, slot sizing and commitments for BigQuery, and cluster and warehouse sizing for Databricks.
Fixes your data team can ship
Ranked changes, from partitioning and clustering to auto-termination and query rewrites, each with its estimated saving.
What is included
Every package includes all of this. The tiers below change the size of the estate we cover and how much hands-on help you get.
BigQuery query and job analysis
30 days of job history per project: the most expensive users, service accounts and repeated query patterns, and what each costs.
BigQuery slots and editions
On-demand versus Standard, Enterprise and Enterprise Plus editions, slot sizing from your real usage percentiles, idle reserved slots and commitment options.
BigQuery tables and storage
Cold and unused tables, queries missing partition filters, missing partition expiration, partitioning and clustering candidates, and the logical versus physical storage billing choice per dataset.
Databricks compute
Cost by SKU, workspace, cluster, job and SQL warehouse; all-purpose versus jobs compute; idle clusters and auto-termination; right-sizing, spot and Photon usage.
Databricks warehouses and serving
Serverless versus Pro SQL warehouses, Model Serving and Vector Search costs, and storage compaction opportunities.
Guardrails
The quotas, maximum-bytes-billed limits, budgets and alerts that stop the next runaway query or forgotten cluster.
What we hand over
- Findings workbook (Excel) ranked by saving, with owners
- Pricing model recommendation: on-demand, editions, slots or cluster sizing
- Rewritten SQL for your costliest query patterns (Standard and Complete)
- Guardrails checklist with the exact settings to apply
- 60-minute readout with your data platform team
How it works
Day 0
Free scoping call
Thirty minutes on your platforms, projects or workspaces and goals. You get a written quote that matches a package.
Day 1
Connect read-only access
For BigQuery, viewer roles granted to a dedicated Finitizer reader identity per project. For Databricks, a service principal that can read system tables.
Days 2 to 8
Analysis
We analyze 30 days of history, then a practitioner reviews every finding and prices it.
Days 8 to 10
Fixes and pricing plan
Ranked fixes, the pricing model recommendation and, on Standard and Complete, rewritten SQL for your costliest patterns.
Days 10 to 14
Readout
A walkthrough with your data team, then email follow-up while you ship the first changes.
Pricing
One fixed fee per audit, agreed before we start. Paid by card or by invoice at kickoff. Prices are in US dollars and exclude applicable taxes.
Focused
One platform: up to 5 BigQuery projects or 1 Databricks workspace
Typical spend: $5,000 to $25,000 a month
$2,500USD, one-time
- Full query, storage and compute analysis
- Findings workbook ranked by saving
- Pricing model recommendation
- 60-minute readout call
Standard
One platform: up to 20 BigQuery projects or 3 Databricks workspaces
Typical spend: $20,000 to $100,000 a month
$5,000USD, one-time
- Everything in Focused
- Rewritten SQL for your 10 costliest query patterns
- Slot and commitment plan, or cluster policy recommendations
- Guardrails checklist with exact settings
- 30 days of email follow-up
Complete
BigQuery and Databricks, or up to 50 projects or 6 workspaces
Typical spend: $75,000 a month and up
$7,500USD, one-time
- Everything in Standard
- A working session to apply the guardrails with your team
- A second analysis after 30 days showing what changed
- Larger estate? Book a call and we will quote it.
- Start FinOps-as-a-Service within 60 days and the full audit fee is credited to your retainer.
Twice the fee, or your money back
If the audit does not identify at least twice its fee in annual savings, estimated conservatively from your own usage data, we refund the fee in full. It is written into your order, and you keep the workbook either way.
Access and security
We read job, query and cluster metadata. We never read the data in your tables, and we never delete anything.
- BigQuery: viewer roles granted per project to a dedicated Finitizer reader identity, using short-lived impersonation tokens. No keys are created, and no organization-level role is needed.
- BigQuery Resource Viewer and Metadata Viewer are required. Resource Admin, used read-only for slot timelines and commitments, is optional; without it the slot analysis is less detailed.
- BigQuery job history includes query text and the user who ran each job. We use them to attribute cost, keep them in a dedicated dataset for your company, and delete them on request.
- Databricks: a service principal with SELECT on the system tables and permission to use one SQL warehouse. No workspace admin, cluster or job permissions. Revoke its token when we finish.
Cannot grant access yet? For BigQuery, run the free BigQuery Cost Check at tools.finitizer.com. You run three read-only queries yourself and the report is computed in your browser; we can start the conversation from its results.
Who does the work
Nitin GandhiFounder, Finitizer
Every engagement is led by Finitizer's founder, a FinOps practitioner who has run cost programs inside large enterprises, not just advised them. The Finitizer platform does the scanning; the judgment is human.
- 25 years in enterprise IT
- FinOps programs for IT budgets of more than $220 million
- Five FinOps certifications
- MBA, Cornell University
Frequently asked questions
Do you read the data in our tables?
No. We read job and query metadata, table metadata and cluster usage. BigQuery job history does include query text and user names, which we use only to attribute cost.
Will your analysis increase our BigQuery bill?
Barely. We read INFORMATION_SCHEMA job and storage views, which process a small amount of metadata; for most estates the whole audit costs a few dollars of query time.
Will we see savings before the audit ends?
Often, yes. Safe quick wins such as expiring old partitions, switching a dataset to physical storage billing or adding auto-termination are flagged as soon as we find them.
Should we move from on-demand to BigQuery editions?
It depends on how steady your usage is. We size slots from your real usage percentiles and show the monthly cost of each option side by side, including commitments, so the choice is a number rather than a guess.
Do you need organization-level access in Google Cloud?
No. Roles are granted per project, and we read job history project by project.
What does Databricks need on its side?
Unity Catalog with system tables enabled, and a service principal with read access to the system schemas and use of one SQL warehouse. Nothing else.
Our costs come from Looker, dbt and BI tools. Can you still attribute them?
Yes. We group work by service account, labels and repeated query pattern, so a dashboard or a dbt model shows up as one line with its own cost.
Will you change our reservations, clusters or queries?
No. We recommend and, on Standard and Complete, rewrite the SQL. Your team applies the changes. On Complete we work alongside you while you set up the guardrails.
Is Databricks part of the Finitizer platform?
Databricks is delivered as this audit, using our Databricks analysis engine. Ongoing monitoring in the platform is available for BigQuery today.
Will you sign an NDA and complete our security questionnaire?
Yes. We sign your NDA before any access is granted, send the exact roles we need in advance, and complete your security questionnaire. Our SOC 2 program is in progress.
Do you audit Snowflake?
Not today. If you are comparing warehouses, the free warehouse cost translator at tools.finitizer.com may help.
Ready to start?
Pick a package above, or book a free 30-minute call and we will tell you which one fits, or whether you need one at all. From $2,500, one-time fee.