FinOps strategy development and execution

“FinOps is an operational framework and cultural practice which maximizes the business value of cloud, enables timely data-driven decision making, and creates financial accountability through collaboration between engineering, finance, and business teams.” (FinOps Foundation)

FinOps is an operational framework and cultural practice that, through collaboration among engineering, finance, and business teams, enables timely data‑driven decision‑making, establishes financial transparency and accountability, and maximizes the value of cloud business.

Establish cost optimization process

The FinOps Framework presented by the FinOps Foundation outlines the principles, domains, capabilities, stakeholders, and stages for cost optimization as shown in the figure below.

To understand the core concepts and processes of cloud cost optimization, let’s examine the elements of the FinOps Framework.

FinOps Framework: Principles(Principles)

The principles presented by the FinOps Framework are as follows.

• Team Collaboration

  Since the cloud incurs costs per resource and per hour, finance, engineers, product/service, and business teams must collaborate continuously. Each team continuously collaborates to improve efficiency and drive innovation.

• Decision-making based on the business value of the cloud

  Unit economic efficiency and metrics better reflect their impact on the business than total spending. Consider trade-offs among cost, quality, and speed. I consider the cloud to be a driving force for innovation.

• Assign responsibility for cloud usage to all entities

  Responsibility for usage and cost extends to lower-level organizations or business units, and engineers are accountable for costs from architecture design through current operations. Each functional team and product team manages cloud usage within their department’s budget limits. Distribute decision-making on cost‑effective architecture, resource utilization, and optimization to each department. The technical team should consider cost as a new efficiency metric from the early stages of the software development lifecycle.

• Real-time viewable FinOps Data

  Cost data is processed and shared immediately upon creation. Ensuring real-time data visibility increases cloud utilization. Rapid, continuous feedback on data contributes to generating efficient results. Information about cloud spending costs is provided to all departments of the organization in real time. Generate, monitor, and improve real-time financial forecasts and plans. We conduct trend and variance analysis to understand the causes of cost increases. We promote best practices and celebrate achievements through internal team benchmarking. We evaluate our performance by benchmarking against peer organizations in the industry.

• FinOps led by dedicated team

  While each department holds responsibility for its own area, the dedicated team plays a role in recommending, disseminating, and supporting best practices. The executive team supports FinOps and its practices and processes. We integrate management of fees, contracts, and discount optimization to leverage economies of scale. Engineers and the operations team focus on optimizing usage in their respective environments and do not get involved in discount negotiations.

• Utilizing the cloud’s variable cost model

  I view the cloud’s variable cost model as an opportunity to provide greater value rather than a risk. We predict, plan, and purchase capacity just-in-time when needed. We establish agile iterative plans rather than static long‑term plans. Do not let irregular pushes cause you to tidy the system; achieve cloud optimization through proactive system design and continuous adjustments.

FinOps Framework: Stakeholder(Persona)

To implement FinOps, various stakeholders within the organization, i.e., personas, must work together with the FinOps team.

FinOps activities are not performed only by the FinOps team.

Everyone involved in the use, tracking, management, or direction of cloud utilization is a FinOps stakeholder, and through collaboration within the FinOps Framework, each gains benefits.

The following table shows the core FinOps stakeholders, associated organizations, and their roles.

FinOps Framework: Phase (Phase)

FinOps performs framework tasks by repeatedly executing the three stages of Inform(notification), Optimize(optimization), and Operate(operation).

Inform: Visibility and Allocation

In the Inform stage, activities are performed to collect and analyze data on cloud costs, usage, and efficiency.

In this stage, the tasks include identifying data sources, allocating and analyzing data, and then preparing a report.

Through this, each team can gain the capability to develop metrics that enable budgeting, trend forecasting, KPI establishment, and explaining the business value of cloud spending.

By accurately allocating cloud spend according to tags, users, or business rules, precise reporting becomes possible, and the business and finance teams can increase ROI within budget while forecasting spend and carbon costs.

Cross-team benchmarking provides a useful metric for evaluating an organization’s operational efficiency.

Additionally, by linking cloud cost data with sustainability, efficiency, utilization, and performance benchmarks, you can derive key performance indicators (KPI) and unit metrics for the organization’s cloud usage.

The cloud offers on-demand access, elasticity, and various pricing discounts, and organizations must build a data‑driven decision‑making framework on this foundation, monitor spending in real time, and continuously communicate business objectives to each department.

Optimize: Pricing and Usage

In the Optimize stage, opportunities to improve cloud efficiency are identified by leveraging the data and capabilities gathered during the Inform stage.

In this stage, you use the resource optimization tools provided by the cloud provider to adjust underutilized resources to appropriate sizes, apply the latest architecture, manage workloads, and eliminate unnecessary resource waste.

Additionally, visibility, analysis, and reporting features for strengthening the purchase and management of pricing models such as contract discounts and Cost Savings are also included.

In this stage, areas where cloud performance does not align with the organization’s value goals are identified, and cross‑team collaboration is carried out to optimize reporting and management processes based on unit metrics.

Conflicts may arise between optimization directions, but the ultimate goal is for the organization to generate greater value through cloud investment.

Operate: continuous improvement and usage

In the Operate stage, we implement changes to embed FinOps within the organization based on the data and capabilities secured during the Inform and Optimize stages.

This stage includes establishing cloud governance policies, monitoring compliance, developing training programs aligned with organizational goals, establishing team guidelines and automation policies, and granting individual permissions.

For FinOps to succeed, engineering, finance, and business teams must collaboratively take continuous, incremental actions based on the data generated in the Inform stage.

Additionally, you must select the optimal opportunities identified in the Optimize phase and build a culture of accountability for the actions implemented across the organization.

In the Operate stage, we set goals to repeatedly develop strategies and improve workflows, and when necessary, return to the Inform and Optimize stages to mature activities and introduce new capabilities.

Through this, you can continuously improve the organization’s FinOps operations.

FinOps domain and activities

The domains of the FinOps Framework are the fundamental business outcomes that an organization must achieve through FinOps, and FinOps activities refer to the tasks performed within each domain.

The first domain is about understanding cloud usage and costs.

In this area, we collect and analyze cost-related data, detect anomalies, and manage them.

The next domain is business value quantification.

In this area, we quantify the potential costs and value of the business and conduct activities that analyze economic feasibility through benchmarking.

The third is the domain of cloud usage and cost optimization.

In this area, we measure the usage of the cloud workloads being performed and develop and implement strategies to optimize costs.

The last is FinOps Execution Management.

We establish a culture and organization that execute cloud strategy within the organization through FinOps strategies and processes.

FinOps maturity

FinOps activities are performed repeatedly, and through this, the maturity of processes, domains, and activities improves.

The FinOps maturity approach starts small and gradually expands in scale, scope, and complexity.

By taking swift, necessary actions within a small, limited scope and performing FinOps, you can later expand FinOps activities in a larger, faster, and more granular manner based on the results of those actions.

The table below shows the maturity of FinOps.

Cost Requirements Analysis

Based on the FinOps execution stages examined earlier, tasks related to cost requirements can be derived.

The ultimate goal is to reduce total costs through cost optimization, but tasks and task objectives can be set for each implementation phase to achieve this.

The following outlines the requirements for each execution stage.

Inform stage

Optimize stage

Operate phase

Cost Modeling

In an on-premises environment, demand is forecasted, and fixed costs are allocated in the early stages to build the information system.

In contrast, in a cloud environment, resource costs are incurred based on actual demand, and the costs can vary on a monthly, daily, or hourly basis.

Therefore, using the cloud improves efficiency and agility, but the variability in cost and usage also increases.

When customer demand increases, resources also expand, leading to higher costs, but if efficiency improves, the cost increase can be reduced.

The cloud’s flexible nature also affects cost and forecasting.

Develop plans that reflect changing business requirements and demand, and incorporate this volatility into the existing organization’s budgeting process.

To forecast the cost of a new workload, cost modeling should be taken into account.

Cost modeling is the process of setting goals with FinOps stakeholders through financial analysis based on an understanding of cloud costs, and identifying cost‑optimization opportunities.

Methods for assessing volatility for forecasting include Business Practice and Data Analytics approaches, and the items to be reviewed for each approach are as follows.

• Business Practice Approach

  • Identify periods of periodic user increase such as peak seasons and holidays.
  • Future plans that can increase users, such as projects and marketing events
  • Data collection cycle for forecasting

• Data Analytics approach

  • Verify whether past data reflecting periodic user growth is incorporated
  • Verify whether past data is reflected in the business event
  • Gap analysis between predicted and actual results
  • Long-term trend forecast

Next, we need to segment the data to define the cost.

The level of granularity can vary depending on the organization’s structure and internal processes, as well as specific requirements and The granularity level is determined according to the organization’s circumstances.

Using this, you can conduct cost modeling based on the following example.

Group all expenses according to the classification criteria.

Calculate the average spending for the past three months by group.

• Calculate the average value of the same period one year ago.

Compare the two averages to identify the trend and apply the growth rate to each group.

Reflect all future plans.

Share with the FinOps stakeholders and coordinate feedback.

CiO-based cost management

CiO is Samsung SDS’s collaboration platform to support FinOps processes.

This platform supports real-time information sharing, inter‑team communication/collaboration, and immediate execution.

• Enterprise-wide platform for establishing FinOps Providing information, execution guides, and processes for Hybrid, Multi-Cloud deployment/operation.

• Activating collaboration through real-time cloud information sharing and communication Verify the purpose, method, and scale of resource usage. Verify how much effort the development/operations department is putting into resource optimization. Check how much cloud costs are being reduced at the enterprise level.

• Providing platform value for cost discount and resource optimization Usage pattern-based fee discount and resource optimization recommendation/execution (Short-term) Additional profit sharing (Long-term) of the Enterprise Discount Program through integrated resource management

• Providing FinOps standard processes and continuously updating based on execution Traditional resource management process → Establishing an optimal collaboration process for cloud environments

CiO provides concrete execution strategies and collaboration platforms to advance FinOps to the ‘execution’ stage.

Providing only actual usage and cost information presents many challenges in executing FinOps.

Accordingly, CiO defines management metrics and processes for performing FinOps and supports their execution.

Through FinOps management metrics, you can intuitively and quantitatively understand the status of cloud resources and cost management.

The results of each metric item are linked to improvement activities, enabling continuous governance execution.

It is necessary to define the management level based on the organization’s current status and to systematically carry out change management.

For efficient decision-making, a collaborative process among FinOps participants is essential.

The above diagram shows the process established by applying FinOps’s basic process of Inform → Optimize → Operate.

To ensure that the process defined in this way can be continuously executed and managed, systematic systematization is required.

The CiO of Samsung Cloud Platform is one of Samsung SDS’s Managed Service offerings, and for more details, please refer to Samsung SDS Cloud in One.