Organisations need to take a structured and proactive approach to predicting, tracking and managing the cost of their cloud services to meet their financial objectives, writes Jon Bryant CITP MBCS, Consultant with Strategic Blue.

Cloud FinOps is the financial management processes and practices for cloud services. Its core purpose is to maximise the value of cloud services for your business. Modern cloud operations provide the opportunity for great flexibility and cost-efficient delivery of services, but also the potential risk of uncontrolled spending. Cloud FinOps is not simply another name for cloud cost optimisation, but has a much wider scope across the business.

Why do we need cloud FinOps?

Historically, organisations have bought servers, storage and networks regularly as a significant capital expenditure. This was agreed and approved between IT management and key business stakeholders who needed compute services. Procurement would then negotiate and buy the hardware and software. The IT team would have them installed and configured. The finance team would have clear visibility of the purchase process and could allocate the costs to an appropriate budget. If it was charged back to individual departments or teams then the allocation was normally agreed on an annual basis.

Due to the ‘on-demand self-service’ and ‘rapid elasticity’ nature of the cloud and the rate of technical and business change that can be enabled by cloud services, the cost of cloud can be highly dynamic and will need regular review and adjustment to maintain an accurate forecast. There are multiple potential drivers for change, both direct and indirect actions, such as:

  • News and social media articles can generate a lot of cloud traffic leading to a significant leap in costs.
  • A successful sales and marketing campaign in one area can significantly change cloud costs in another area.
  • Quarterly and annual events can massively impact on your cloud costs.
  • Small features changes can have a disproportionate impact on the cloud costs.
  • Engineers can deploy new servers, services or whole environments almost instantly, which all have a cost.

For you

Be part of something bigger, join BCS, The Chartered Institute for IT.

This capability for multiple teams to impact upon the cloud costs can raise significant questions and create frustration within a business, with a perceived lack of control at the core. Cloud FinOps is a way to address these dynamic events by driving and delivering a cultural change to cloud cost management throughout the business. All the stakeholders for cloud services need to be aware of, and involved in, the delivery of cloud FinOps to ensure that a coordinated approach is taken.

Maintaining a balance

In the dynamic cloud environment, cloud FinOps can facilitate the communication between the business, finance and technical teams to help answer the key questions which help to define the business needs for the cloud. These questions can include:

  • Performance v cost: How much am I willing to pay to provide the performance I need? What is the limit to the scalability that should be built into a service?
  • Resilience & availability v cost: How robust does the solution really need to be? Does resilience need to be built across zones and/or regions? How much data can be lost (RPO)? What is an acceptable level of downtime for the business (RTO)?
  • Ease & familiarity v cost: How much are you willing to change your design and operational processes as you move to cloud services? Do you really need the dev and test environment 24x7 or just normal working hours? Can you implement infrastructure as code to deploy and remove services quickly?
  • Control v flexibility: How do you avoid large overspend risks while enabling the engineering teams to work flexibly and effectively?
  • Engineering time v savings: Can you redevelop parts of the application to be serverless functions? What is the ROI on the engineering effort needed to minimise the cloud costs? Can I afford engineering time that isn’t delivering new functionality and features? What should I automate and what is too unique to make automation cost-effective?
  • Flexibility v savings: What can I commit to without sacrificing business and technical flexibility to access discounts? Is the service time-critical or can I use an interruptible solution?
  • Vendor solution v cost: Is this the right cloud provider for this solution? Which country does the data need to be in? Can this service be in a more cost-efficient region?
  • Value of FinOps v Cost of FinOps: Does the FinOps team justify its cost? What non-financial measures should also be used to show the value of Cloud FinOps?

As you can see from these questions, cloud cost is now an actively managed part of the solution and effectively an input into cloud design rather than just an output of the design process.

Stakeholder Coordination

You should establish a central cloud FinOps function to provide coordination and information between the many stakeholders who are involved in cloud operations. This function will normally own the cloud service provider relationships and may be responsible for negotiating discounted rates in exchange for long term cloud spend commitments. Key considerations could be:

  • Executive leadership: Need to know that the cloud is controlled and managed effectively and efficiently. They should be looking at the value of cloud and its impact on overall profitability.
  • Finance team: Need to identify the budget owners, forecast the costs for each budget owner, and monitor the actual v predicted costs on a regular basis. An understanding of the variability of different workloads helps comprehension of changes.
  • Product owners: Need to forecast and track the cloud costs related to their product, and account for those costs within the P&L. They need to understand the business changes that can impact their predicted costs, and track value metrics to ensure they are getting an efficient service.
  • IT leadership: Need to ensure that the cloud FinOps strategy is incorporated into the cloud strategy, and ensure an effective cloud FinOps service. They need to establish optimisation priorities, define technical controls and processes so that they can track the trends, identifying and responding to any anomalous costs.
  • Engineering teams: Need to be aware of their service costs and have defined routes for the technical optimisation of their services. They need to be alerted to anomalous or non-forecast costs and either take corrective action or communicate the changes and their value. The running cost needs to be considered as part of the design process when developing or refreshing systems to enable optimum design decisions.

What is the Cloud value?

Throughout this article the concept of cloud value has been mentioned several times. The vision for a mature cloud FinOps practice is to be able to move away from cost tracking and articulate the value of the cloud through what is known as the unit economics, such as cost per transaction or cost per user or usage hour.

These metrics provide a way for you to track the cloud business value over the long term. For example, even after significant cloud FinOps effort the cloud costs may have doubled but if in the same period customer numbers have tripled, the cloud value has improved and your FinOps practice can demonstrate its value of the FinOps foundation.