Getting Cloud FinOps Right

Bob Scheier: What changes are you seeing in cloud adoption and spending?

Nikhil Roychowdhury: Over the last few years there’s been an immense acceleration to the cloud by our clients, across different industries, moving large portions of their infrastructure and applications to the public cloud. Some of this was driven by COVID. What we’re seeing now, as a result, is cloud becoming a material expense that has gotten more attention from CFOs and CIOs, who maybe did not have the expectation their cloud spending would be as high as it is.

How have organizations responded?

There is a recognition that there needs to be some sort of discipline around how we manage and track this cloud cost. What are the levers that would help optimize that spend? This has given rise to the discipline of FinOps, bringing finance, technology, and engineering to the table with the business to discuss how to bring accountability to cloud spending.

What is different about cloud infrastructure spending that requires something like FinOps?

There have been misconceptions that cloud is always the cheaper option compared to on-premise infrastructure. That is not always the case, especially when workloads are moved “as is” to the cloud without modernization. The economics of the cloud are fundamentally different from how things were done when a company had to purchase its own hardware. The cloud shifts IT spending from the traditional cycles of heavy capital spending on hardware and software to an ongoing operating expense.

In the cloud, the ease of provisioning infrastructure means there is much more potential for hidden or wasted cloud spending as business units, departments or even individual users easily purchase cloud capacity to meet their immediate business needs without going through a traditional capital expenditure process. FinOps attempts to provide visibility into this spend, to ensure waste is minimized, and to link it back to business value. 

How good a job are most companies at FinOps?

We see clients at different levels of maturity. For those just starting to get their feet wet with public cloud, we advise them to establish a FinOps capability before they bring their first application to the cloud. That includes bringing the business, technology and finance functions together to work on how they are going to plan for, and continually optimize their use of the cloud. We might classify these companies as being in the “crawl” stage before they graduate into the “walk” and eventually “run” phase.

Some of our other clients who already have a large portion of their applications in the cloud, tend to be further along in using FinOps to optimize their cloud spending. These clients are thinking about more advanced cloud use cases such as generative AI or broadening their FinOps capability beyond cloud to their on-premise infrastructure.

Where do companies go wrong with FinOps?

One major mistake is to think that only one function, such as finance, or technology, can drive FinOps without bringing other stakeholders to the table. We find such localized FinOps efforts tend to focus too heavily on governance and cost reduction and limit the adoption of FinOps practices. The most successful FinOps efforts bring together stakeholders from the CIO to the CFO and business unit leaders to monitor and manage enterprise cloud spending with a focus on driving business value. If any of those stakeholders are not involved, we tend to see a lack of FinOps adoption, and a lack of understanding of how FinOps can make the most effective use of IT cloud spend. 

For example?

The FinOps team may think they have identified an opportunity to save money by moving a set of servers or data storage to a lower performance tier with the cloud provider. The technologists may be very happy with the change, and the finance function is very happy to have found the potential savings.

However, when you engage the business teams, they may have a very valid reason for why downgrading the infrastructure won’t work. The higher performance compute or storage may be needed to support a new product or service that will drive revenue. Or over provisioning might be required for regulatory reasons, such as to ensure certain levels of uptime or performance. Not having that upfront engagement with business teams can give FinOps a bad reputation throughout the enterprise and make it harder to expand its legitimate use.

How can CIOs help to implement FinOps most effectively?

What I’ve seen be most successful is a “crawl, walk, run” approach. The first step, crawling, is when the different parts of the business and various stakeholders recognize the value FinOps can deliver. In this stage the CIO can help by speaking about the benefits of FinOps, and creating a small, nimble team to drive this awareness.

Engagement methods such as town halls or “lunch and learns” can be highly effective in educating business and technology leaders about how to incorporate FinOps within cloud programs. This also helps with change management – educating all stakeholders about how FinOps can help them do their jobs and reassuring them it is not a threat.

In the crawl stage, FinOps teams are usually establishing basic capabilities like ensuring detailed visibility and reporting into cloud spend and allocation, and engaging with business teams to understand their needs and how the central FinOps team may help them. In this stage, companies may develop a charter and a plan for how their FinOps capabilities will evolve and mature to benefit the business over time.

In the “walk” stage, multiple FinOps capabilities may have already been implemented, cost-efficiency actions have been executed, and teams have started publicizing small successes, showing how FinOps can both save money and grow the business. In this stage, organizations may have started building or planning for advanced capabilities such as automating certain advanced use cases. FinOps is well established and business teams have started realizing its value.

The “run” stage is where an organization truly starts to have an enterprise-wide impact. During this stage, driving more automation is at the core of FinOps processes and the team starts focusing on more advanced use cases such as relating cloud spend to business metrics, or even expanding FinOps capabilities outside of public cloud to encompass the on-premise landscape.

How big should the FinOps team be and who should it report to?

There is no single formula or one-size-fits-all approach. There are many factors at play, including the size of the business, the operating model and overall cloud spend. The point is to not create a big bureaucratic organization or an additional governance layer. The philosophy should instead be one of enablement, allowing business and product teams to be successful and thrive in the cloud. It’s not just about cutting costs; it’s also about creating value.

Where can an organization turn for help getting started?

The nonprofit FinOps Foundation is a great community resource for practitioners. It provides a framework for FinOps and offers a number of certifications as well as training material and other resources.

Final thoughts?

You’re doing FinOps wrong if it becomes a “department of no” focused only on cutting costs, regardless of the effect on the business. Effective FinOps is about enabling business, but in the most efficient way. 

Nikhil Roychowdhury is a principal with Deloitte Consulting’s Financial Services practice, with more than 20 years of experience advising global banks and insurance providers. He leads platform and application teams to deliver scalable cloud platforms and enable new use cases made possible by cloud.