FOCUS 1.2 and 1.3 explained: the cloud billing standard

Every cloud provider invoices in its own dialect. AWS's Cost and Usage Report, Azure's cost exports and GCP's BigQuery billing tables disagree on column names, on what "cost" means, on how amortization is represented, and on where discounts hide. Anyone who has built a multi-cloud chargeback pipeline has written — and maintained, forever — the same brittle translation layer.

FOCUS (FinOps Open Cost and Usage Specification), stewarded by the FinOps Foundation, is the industry's answer: one schema, one set of semantics, emitted natively by the providers themselves. And it has crossed the adoption threshold where ignoring it is the risky choice: per the State of FinOps 2026, 68% of organizations spending $100M+ already use or pilot FOCUS data, and the Foundation reports 93 of the Fortune 100 using it.

The core idea: four cost measures, not one

The single most useful thing FOCUS standardizes is that "cost" is four different questions, each with its own column on every billing row:

The gaps between the four are where FinOps lives: ListCost − ContractedCost is your negotiation, ContractedCost − EffectiveCost is your commitment program, and BilledCost vs EffectiveCost is the amortization timing difference that makes finance and engineering argue about "the real number". With native exports, an Effective Savings Rate or a cost-per-customer metric is computed identically on AWS, Azure and GCP — no translation layer.

FOCUS 1.2: SaaS joins the schema (ratified 29 May 2025)

FOCUS 1.2, generally available since June 2025, is the "Cloud+" release — it extends the same schema beyond IaaS (announcement). The headline additions:

• SaaS and PaaS billing in the same schema as cloud — your Datadog, Snowflake or Databricks line items normalize next to EC2.
• InvoiceId — every cost line links to the invoice that billed it, closing the loop between cost analytics and accounts payable at month close.
• Virtual currencies, credits and tokens — new columns for vendors that bill in credits or tokens (a direct nod to AI-era billing), including burndown tracking and conversion to real currency, plus national-currency support.
• Refined billing account / sub-account granularity and an extensibility model: provider-specific columns carry an `x_` prefix, so custom data is identifiable instead of ambiguous.

FOCUS 1.3: commitments become data (ratified 5 December 2025)

FOCUS 1.3 — the fifth major release, announced 11 December 2025 — targets the places where billing data was still opaque:

• Split Cost Allocation columns. When a provider splits shared costs (Kubernetes pods on a shared node, shared database instances), it must now expose the *methodology* used, not just the result. You can finally audit how your container costs were allocated.
• Contract Commitments dataset. A separate, supplemental dataset isolating commitment terms from cost lines: start/end dates, remaining units, descriptions. One query returns every active RI/SP/CUD with its expiration — the data you previously scraped from three consoles to manage renewals.
• Recency and completeness metadata. Each export carries a last-updated timestamp and a completeness flag, so pipelines can distinguish final data from data still subject to revision — and stop processing stale exports as truth.
• ServiceProviderName / HostProviderName. Who sells you the service vs where it runs — disambiguating marketplaces and resellers. The old Provider and Publisher columns are deprecated, with removal planned in 1.4.

FOCUS 1.4 is already in development (started December 2025), with an Invoice dataset, extended Contract Commitment dimensions, and cross-provider consistency requirements on the announced roadmap.

Who actually supports it in 2026

How to start using FOCUS this quarter

1. Turn on the native exports. AWS: create a Data Export with the FOCUS 1.2 table (this is separate from any legacy CUR you already have — run both during transition). Azure: add a FOCUS dataset to your Cost Management exports. GCP: configure the FOCUS BigQuery view per the official guide. OCI: enable the FOCUS option in Data Exports.
2. Validate before you trust. Run a one-month reconciliation: sum of BilledCost in the FOCUS export vs the actual invoice, per billing account. Then spot-check the columns your metrics depend on (EffectiveCost populated on commitment-covered usage? `x_` columns documented?). This is where the conformance gaps surface.
3. Rebuild one metric, not the whole stack. Pick the metric with the worst cross-cloud drift — usually showback or savings reporting — and re-implement it on EffectiveCost/ListCost. Keep the legacy pipeline running in parallel for a quarter and publish both numbers with the delta.
4. Ask your vendors the 1.3 question. Tools and SaaS providers that bill you: do they emit FOCUS? Tools that read your costs: do they ingest 1.2 natively and have a 1.3 roadmap (Contract Commitments dataset, recency metadata)? The answers sort the field quickly.

What this changes for practitioners and tools

For practitioners, FOCUS converts billing-pipeline work from translation to validation: ingest the native FOCUS export, verify conformance on the columns you rely on, and build every metric — chargeback, unit economics, ESR, anomaly detection — on EffectiveCost semantics that hold across providers. The FinOps Certified Practitioner curriculum now includes FOCUS, and tooling RFPs increasingly ask for it by name.

For tools, it is becoming the entry ticket. The 2026 evaluation question is no longer "can you ingest AWS CUR?" but "are you FOCUS-native — can you ingest 1.2 exports from all my providers *and* handle 1.3's commitment dataset and recency metadata?" A tool that normalizes into its own proprietary schema is rebuilding the problem FOCUS just solved.

Cloudios is built FOCUS-aligned: real billing ingestion from AWS, Azure and GCP normalized to FOCUS semantics, EffectiveCost-based chargeback and ESR, and commitment expirations tracked as data rather than calendar reminders. If you want to see what your bill looks like through that lens, the live demo is open, and pricing is public — both without a sales call.