FinOps: A Personal Story of Cloud Cost Optimization and Practical Steps
This article introduces FinOps, explains its definition and cultural significance, illustrates the concept with a personal household billing story, and outlines five concrete steps—cost allocation, target setting, budget control, operational management, and intelligent automation—to achieve cloud cost optimization.
FinOps is becoming increasingly popular, with Google search trends showing a steady rise since 2019. The CNCF 2023 Cloud Native Trends report lists FinOps, GreenOps, GitOps, and cost reduction as four of the ten hot topics.
What is FinOps? According to the official definition, FinOps establishes a set of cloud financial management rules and best practices that enable engineering, finance, and business teams to collaborate on data‑driven cost decisions, maximizing organizational value.
FinOps is essentially a cultural practice that involves mindset shifts, standardized processes, and team collaboration.
Personal Story: The author uses a household expense bill to illustrate FinOps. In January, the family spent 1,380.5 CNY on living costs out of a 5,000 CNY income, prompting a goal to reduce February expenses to 500 CNY, saving 880.5 CNY.
FinOps Steps
Step 1 – Cost Allocation
The bill shows only total costs without sub‑item breakdowns, so the first step is to categorize each expense. Since mature estimation methods are lacking, the author uses rough percentages based on experience.
Step 2 – Define Optimization Targets
After understanding cost distribution, the author identifies high‑impact items (floor heating, charging station, air‑conditioning) and proposes specific actions:
On‑Demand Start/Stop: Turn off floor heating on warm days, saving ~600 CNY.
Off‑Peak Usage: Charge the electric vehicle at night at half price or use free swapping, saving ~270 CNY.
Usage Optimization: Set air‑conditioner to energy‑saving mode and lower temperature, saving ~26 CNY.
Total estimated savings for this step: 896 CNY.
Step 3 – Budget Control
Based on the previous month’s bill and the optimization plan, set budgets and quotas for each expense category. Two control methods are discussed:
Pre‑Control: Switch to prepaid billing with daily caps (rejected due to risk of power outage).
Post‑Control: Use smart devices to report real‑time usage to a cloud‑native monitoring app, set alerts, and adjust consumption promptly.
Step 4 – Operational Management
High‑level authorization, team collaboration, and individual accountability.
Continuous optimization to keep top‑cost items under control.
Quota management and resource request workflow.
Unit economics with delegated financial authority.
Commitment discounts to lower unit costs.
Resource pricing incentives (daytime vs. night‑time electricity rates).
Step 5 – Intelligent Optimization
Automation tools can realize continuous cost savings:
Cost Visualization & Allocation: Smart appliances record per‑person usage, enabling detailed reports.
Quota Management: Enforce maximum usage per person.
Usage Optimization: Automatic scaling for floor heating, auto‑switch to energy‑saving mode for air‑conditioning, and time‑based charging for electric vehicles.
Result: By applying the above steps, the author reduced March household expenses to 220 CNY.
FinOps combines financial management, operational governance, and cost control, requiring automation to be effective. Applying the same methodology to cloud billing—known as Cloud FinOps—helps organizations manage complex, multi‑dimensional cloud cost structures, involving numerous services, pricing models, and cross‑functional teams.
For deeper insight, the article promotes the book Cloud FinOps (Chinese edition: 《FinOps云成本优化》) published by Electronic Industry Press, available on JD.com, and mentions the open‑source project Crane from the Tencent Cloud Native team.
DevOps Cloud Academy
Exploring industry DevOps practices and technical expertise.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.