Google Cloud FinOps Guide: Labels, Billing Export & Showback

FinOps in Google Cloud, simply explained

FinOps is a way of working where the people who build cloud infrastructure also understand what it costs. Instead of one person checking the bill at the end of the month, every team can see what their services cost in near-real-time.

In Google Cloud, that means setting up billing export so cost data lands in BigQuery, labelling resources so you know which team owns what, and building dashboards that show each team their own numbers. The goal is not to spend less. It is to spend deliberately, with full visibility.

FinOps vs cost optimisation

These terms get used interchangeably, but they solve different problems. Understanding the difference prevents you from skipping the organisational work that makes technical savings stick.

FinOps creates the conditions for cost optimisation to happen continuously. Without it, optimisation is a one-time project that regresses as new waste accumulates. For the technical optimisation actions themselves, see Cost Optimisation Strategies.

How FinOps works in Google Cloud

FinOps in GCP follows a practical workflow. Each step builds on the one before it.

1. Make spend visible

You cannot manage what you cannot see. The first step is getting cost data out of the billing console and into a place where teams can query and explore it.

• Billing account structure. Every GCP resource charges to a project, and every project links to a billing account. Understand how billing works before building attribution on top of it.
• Billing export to BigQuery. Enable the detailed billing export in the Billing console. GCP writes a record for every SKU consumed across every project. This is your single source of truth for cost analysis. It does not backfill, so enable it on day one.
• Cost reports and dashboards. The billing console provides built-in cost breakdown reports for quick overviews. For deeper analysis, query the BigQuery export directly or connect it to Looker Studio for team-facing dashboards.

2. Attribute spend correctly

Visibility without attribution tells you the total but not who generated it. Attribution is what turns a billing report into something actionable.

• Labels. GCP labels are key-value pairs attached to resources. They flow through to billing exports, letting you query cost by team, environment, service, or any dimension you define. Define your labelling standard before creating resources. Retrofitting labels across hundreds of existing resources is painful and always incomplete.
• Project and folder structure. Every cost in GCP is tied to a project. If your projects align with teams or products, you get cost attribution at the project level without labels. Use GCP folders to group projects by team or business unit.
• Team and service ownership. At minimum, every resource should have env, team, and service labels. This lets you answer “which team spent how much on which service in which environment” from a single SQL query.

# Apply labels to a Cloud Run service
gcloud run services update my-api \
  --region=us-central1 \
  --update-labels=env=prod,team=backend,service=api-server \
  --project=my-app-prod

# Apply labels to a VM
gcloud compute instances add-labels my-worker-vm \
  --labels=env=prod,team=data,service=worker \
  --zone=us-central1-a \
  --project=my-app-prod

3. Add guardrails

Visibility tells you what happened. Guardrails tell you before costs go wrong.

• Budgets and alerts. Set billing budgets with threshold alerts at 50%, 80%, and 100% of expected monthly spend. Budget alerts do not stop resources. They only notify. For automatic shutdowns, wire alerts to Pub/Sub and a Cloud Function.
• Cost review cadence. Schedule a weekly or fortnightly cost review where team leads look at their dashboards. Monthly is too slow because a misconfigured resource can run up significant charges in a few days.
• Ownership and accountability. Every project and every significant resource should have a clear owner. When a cost spike appears, there should be no ambiguity about who investigates it.

4. Take action on waste and pricing

With visibility, attribution, and guardrails in place, teams have the data to act on waste and pricing. This is where FinOps connects to the cost optimisation toolkit:

• Rightsize overprovisioned VMs using GCP Recommender data.
• Clean up idle resources such as unattached disks, unused IPs, and stopped VMs still paying for disk.
• Apply committed use discounts and sustained use discounts where workloads are predictable. See GCP Pricing Models for how discount tiers work.
• Use cost investigation tools to find the specific resources driving cost spikes.

For a complete walkthrough of optimisation actions, see Building a Cost Optimisation Strategy.

5. Measure business value

Total spend alone does not tell you whether costs are a problem. A bill that grows from $5,000 to $7,500/month sounds bad. But if revenue grew 60% in the same period, the unit economics are actually improving.

Track cost relative to a business metric:

• Cost per API request served
• Cost per active user per month
• Cost per transaction processed
• Cost per GB of data processed

Unit economics separate healthy growth from growing inefficiency. “This change reduces cost per API request by 22%” is more compelling to a business stakeholder than “this saves $180/month”.

Google Cloud tools that support FinOps

GCP provides native surfaces for each stage of the FinOps workflow. These are the tools you will use most.

Billing export to BigQuery

The foundation of all FinOps analytics in GCP. The detailed export writes every billing line item, including resource-level detail, to a BigQuery table you own. Once enabled, you can query spend by service, project, label, SKU, and time period with standard SQL.

-- Monthly spend grouped by team label
SELECT
  labels.value AS team,
  ROUND(SUM(cost), 2) AS total_cost_usd
FROM `my-billing-project.billing_export.gcp_billing_export_v1_XXXXXX`,
  UNNEST(labels) AS labels
WHERE
  labels.key = 'team'
  AND DATE(_PARTITIONTIME) >= DATE_SUB(CURRENT_DATE(), INTERVAL 30 DAY)
GROUP BY team
ORDER BY total_cost_usd DESC;

Labels and tags

Key-value metadata attached to resources. Labels flow through to billing exports and are the primary mechanism for attributing costs to teams, environments, and services. Define a standard set (env, team, service) and enforce it at resource creation time.

Budgets and alerts

Configurable spending thresholds that send notifications when costs reach a percentage of your budget. Set per-project or per-label budgets to give each team their own spending guardrail.

FinOps Hub

A consolidated surface in the Google Cloud console that brings together cost visibility, committed use discount recommendations, and optimisation suggestions. FinOps Hub aggregates signals from across your billing account and surfaces the highest-impact actions in one place.

Recommenders and optimisation signals

GCP Recommender analyses usage patterns and suggests specific actions: rightsize this VM, delete this unattached disk, purchase this CUD commitment. Active Assist extends this with broader signals across services. Both feed into FinOps Hub.

Looker Studio dashboards

Connect Looker Studio (free) to your BigQuery billing export to build team-facing cost dashboards. Share read-only dashboards with engineering team leads so they can see their own cost data without writing SQL.

When to use this

Not every team needs a formal FinOps practice on day one. But if any of these describe your situation, it is time to start:

• Multiple teams sharing one billing account. Without attribution, no one knows who is responsible for what.
• Costs rising without clear ownership. The bill grows every month but no specific team or service is accountable.
• Frequent cost surprises. Someone gets a bill that is significantly higher than expected and has no idea why.
• Leadership asking for accountability. Finance or management wants to know what cloud money is buying.
• Ad hoc cleanup is not working. Someone does a cost sweep once a quarter, but waste accumulates again immediately. You need a continuous process, not a periodic project.

A practical rollout path

Rolling out FinOps does not require a large team or months of planning. Here is a realistic timeline for a small or mid-sized engineering team.

First week

• Enable billing export to BigQuery on all billing accounts.
• Define your labelling standard: env, team, service at minimum.
• Set budget alerts at 80% and 100% of current monthly spend on each project.
• Label all production resources. Use a script or Terraform update to batch-apply labels.

First month

• Build a basic Looker Studio dashboard showing spend by team, service, and environment.
• Share the dashboard with engineering team leads.
• Run your first cost review meeting: walk through each team’s spend and identify obvious waste.
• Clean up the easiest wins: unattached disks, unused IPs, dev resources left running.

First quarter

• Establish a regular cost review cadence (weekly or fortnightly).
• Add unit economics tracking: cost per request, cost per user, or cost per transaction.
• Evaluate committed use discounts for stable production workloads.
• Review whether showback reports are driving behavioural change. If teams are not engaging, adjust the dashboard or the review process.
• Document the labelling standard and cost review process so it survives team changes.

Showback vs chargeback

Both are mechanisms for connecting cloud costs to the teams that generate them. They differ in how formally the costs are allocated.

Start with showback. It creates most of the behavioural change with a fraction of the setup. Teams that can see their costs typically self-identify waste without needing the pressure of an internal charge. Move to chargeback only when the organisation has mature labelling, a reliable billing data pipeline, and an agreed cost allocation methodology for shared resources.