Skip to main content
Week 1

What Is FinOps, Really?

Read it and work through the exercises — no video required.

~22 min read · Text + interactive

By the end of this lesson, you can…

  • Explain the Inform → Optimize → Operate loop in your own words, without jargon.
  • Match a cost driver on a real bill to the persona who owns the decision.
  • Recognize which of the 6 FinOps Principles is at play in a given cost conversation.

What Is FinOps?

FinOps isn't a tool, and it isn't a job title on its own. It's a discipline for making fast, accurate spend decisions in a cloud where the bill changes every time someone hits deploy. It sits at the intersection of finance, engineering, and product — and it runs on one core loop.

01 · Inform

Same data, everyone

Accurate, timely, shared visibility into what's actually being spent.

02 · Optimize

Act on it

Rightsizing, commitments, waste elimination — the moves data points to.

03 · Operate

Make it continuous

A standing practice, not a quarterly fire drill.

The 6 FinOps Principles

These thread through a large share of the FOCP exam, directly or indirectly (more in Week 8's exam-weighting breakdown). Tap each one.

01Teams need to collaborate

Finance, engineering, and product work together at the speed each function needs.

In practice: a shared Slack channel where engineers see cost anomalies the same day finance does.
02Business value drives technology decisions

Prioritize unit economics over minimizing aggregate spend for its own sake.

In practice: a costlier feature can still be right if it drives proportionally more revenue.
03Everyone takes ownership for their technology usage

Accountability is pushed to the edge — engineers own their architecture's cost impact.

In practice: a team's cost dashboard is as visible as their error-rate dashboard.
04FinOps data should be accessible, timely, and accurate

Cost info gets shared promptly across every level, not gatekept by one team.

In practice: a self-serve dashboard beats a monthly PDF report.
05FinOps should be enabled centrally

A central team evangelizes best practices and negotiates rates, even as ownership stays distributed.

In practice: one team owns the Savings Plan strategy company-wide.
06Take advantage of the variable cost model of the cloud

Lean into consumption-based pricing instead of treating cloud like a static data-center budget.

In practice: scaling down on weekends instead of provisioning for peak all week.

Personas & Motivations

Understand what each stakeholder actually wants. Tap each persona.

Engineering

Will this slow me down?

Watches deploy velocity and how much friction a cost control adds to their workflow.

Finance

Will we blow the budget?

Watches forecast-vs-actual variance and whether spend is predictable month to month.

Procurement

Are we getting the right discount?

Watches commitment coverage and whether existing RIs/Savings Plans still match real usage.

Product

Does the unit economics work?

Watches cost-per-customer or cost-per-transaction as the feature or user base scales.

Leadership

Is this investment paying off?

Watches the headline efficiency trend — cloud spend as a share of revenue, quarter over quarter.

How Cloud Billing Works

AWS, Azure, and GCP price around the same three models. On-demand is the most expensive and flexible. Committed / reserved trades a 1–3 year commitment for a discount. Spot / preemptible is cheapest by far, but the provider can reclaim it with little notice.

Illustrative — same instance type, three ways to buy it

On-demand100%
1-yr commit~65%
3-yr commit~45%
Spot~10–30%

Actual discounts vary by instance family, region, and term — these are typical bands, not quoted prices.

A real invoice blends all three: the same account might run its steady-state database on a 3-year commitment, autoscale its web tier on-demand for traffic spikes, and run overnight batch jobs on spot. That's why "the bill" is really dozens of line items, not one number — and why the same total spend can mean very different things depending on the mix.

Quick glossary: terms you'll see on real bills
CUR
Cost and Usage Report — AWS's raw, line-item-level billing export. The source data behind most cost dashboards.
Tagging
Labeling resources (e.g. team:payments) so cost can be attributed to the team or product that owns it.
Showback vs. chargeback
Showback just reports a team's spend to them; chargeback actually bills it back to their budget.
RI / Savings Plan / CUD
AWS Reserved Instances, AWS Savings Plans, and GCP Committed Use Discounts — different vendors' names for the same idea: commit to a spend level, get a discount.
Unit economics
Cost per customer, per transaction, or per API call — ties infrastructure spend to a business metric instead of a raw dollar total.

Exercise: Read a Real Bill

Flag the top 3 cost drivers and tag each line to a persona, then reveal my analysis.

Line itemCostFlagPersona
Amazon EC2 — On-Demand (m5.2xlarge fleet)$18,420
Amazon RDS — db.r5.4xlarge (Multi-AZ)$9,150
Data Transfer — Inter-AZ + Internet Egress$6,780
Amazon S3 — Standard Storage + Requests$2,340
Idle / unattached EBS volumes$1,240
AWS Lambda — Invocations$410
Reserved Instance amortization (discount)−$4,200

Quick Check: A Different Bill

Same skill, a new scenario. Read it, form your own answer, then reveal to compare notes.

A three-person team ships a new recommendation feature on Tuesday. By Friday, Data Transfer costs for their service have tripled, while everything else on the bill is flat. No one has looked at it yet.

Who gets looped in first, and what question do they actually ask?

Reveal my answer

Engineering, first — not Finance. A sudden, isolated spike right after a deploy is almost always an architecture signal (a chatty new call pattern, an unbounded retry loop, cross-AZ traffic that should've stayed local) before it's a budget problem.

The question isn't "why did this cost so much" — it's "what changed on Tuesday?" That's the Inform phase doing its job: the data surfaced something fast enough that it's still a five-minute architecture fix, not a quarter of accumulated waste. Finance only enters once Engineering confirms whether the new pattern is a bug or the expected cost of the feature.

Reflection

What's one FinOps decision at your company that a persona other than Engineering probably cares about more? If you're not sure yet, guess — who would Finance or Procurement call first if the cloud bill doubled next month?

Putting it together

The Inform → Optimize → Operate loop is the engine; the 6 Principles are the rules it runs by; the personas are who's actually in the room; and the billing model is what makes the numbers move in the first place. Read a bill like you did above, and all four show up on the same page — that's the whole discipline in miniature.

Week 1 of 9 complete11%
Up next

Week 2: Understand Usage & Cost

SQL over billing exports, cost dashboards, tagging & allocation strategy.

That's Week 1.

This is one lesson from the full FinOps Certified Practitioner course.

See the full course →