Why FinOps Must Focus on Value, Not Just Cost
You can’t talk about cloud without talking about cloud costs. When market conditions were healthy it was easy to focus on top-line growth and less on bottom-line cloud cost.
The situation has changed. Cloud cost is a major issue, and it is proven that it is not going to be solved by relying on the tools and reporting provided by cloud providers.
Shock cloud bills are common, and even organizations with a reasonable handle on their cloud usage can struggle to stay within — much less optimize — their budgets: Only 28% of respondents in the latest FinOps Foundation survey said their organization can accurately budget and forecast their cloud costs.
FinOps intends to bring effective financial controls and management to the reality that more and more applications and infrastructure run somewhere outside of the corporate perimeter. If you’re trusting your vendors to do that for you, you’re doing it wrong. That sounds like a good thing, right?
The persistent problem is that FinOps today is too focused on the “Fin” and not enough on the “Ops,” according to David Williams, senior vice president, market strategy at Quali, an infrastructure automation and management platform.
FinOps teams are still playing a reactive game, assessing retroactively where and why costs are too high — with too little visibility into why that cloud usage exists in the first place, and especially how it’s delivering value to the business.
“While most FinOps teams look at having the basic ability to track usage around accounts or cloud instances, that does not always reflect the way that the cloud is used to deliver value to the business,” Williams told The New Stack.
The State of FinOps: Fragmented
It’s not necessarily the fault of those teams. Rather, FinOps in its earliest iterations has suffered from some of the same problems that plagued its predecessors — namely, an approach to cost management that focuses on the “what” — how much the bill says you spent, and only after it arrives — versus the “why,” which should accurately reflect the business reasons for consuming cloud resources in the first place, as well as the results those choices produced.
Moreover, FinOps, even while the name suggests tight collaboration, often still relies on fragmented and retroactive processes and information, according to Williams.
“When the invoice suddenly lands on your doorstep and then you have to work out that invoice — and dozens and dozens of other invoices — with account owners on the fly, it’s always a workaround,” Williams said.
Cost matters, of course, but Williams sees a need for pairing cost with value, and shifting the attention of FinOps away from billing to focus on value at the time cloud resources are provisioned.
“FinOps is about cost accountability, whereas it needs more emphasis on cost and value delivered,” he said. “Value is established when the cloud is provisioned. Not when the invoice arrives.
7 Key Challenges
There are several hurdles organizations typically face when trying to do just that.
1. Lack of Visibility
One of the fundamental problems behind surprise cloud bills is that too many organizations still don’t understand what they are using or running at any given time. The “classic” example here is the developer who spins up a cloud instance for legitimate reasons but then forgets to turn it off.
But the possibilities for unknown (and unnecessary) cloud resource consumption are virtually limitless, especially in large companies or any organization using multiple cloud services.
2. Lack of Accountability
Cloud usage is typically tracked by accounts. While this provides a view of usage by the account owner, it doesn’t capture the purpose of the usage. For organizations with many accounts across different clouds, the task of understanding and assessing why the cloud was used can be a long, manually intense activity that is usually done as each cloud invoice arrives.
Going forward, associating — in real time — the purpose with cloud usage is critical. It is the only way to understand which cloud resources are used, and why. This detail allows organizations to optimize cloud usage (based on purpose value), associate cloud usage with business value, manage cloud usage according to a budget and significantly reduce the time needed to understand cloud usage against the account-based cloud invoicing.
3. Lack of FinOps Readiness and Maturity
FinOps is, in theory, a good thing. But most organizations aren’t yet ready to do it right, according to Williams. There are too many silos and a lack of ingrained culture around cloud discipline, collaboration, and how (and why) cloud resources are provisioned, delivered, and managed.
4. A Focus on Lagging Indicators
Many organizations have focused on their actual invoices — and have even adopted cost management tools that do the same — instead of optimizing costs as they’re accrued, ideally in real-time.
5. Too Much Inherent Complexity and Technical Debt
A decade or more of ungoverned, siloed cloud usage creates inherent complexity and bad habits that are difficult to undo. This is particularly true in terms of how cloud gets used in software development, according to Williams.
6. Lack of Relevant Tools
Too many cloud management tools treat the cloud as an IT asset. In Williams’ view, that’s simply inaccurate — you don’t own your cloud resources, so why manage them as such? “They’re ephemeral and in a constant state of change,” he said.
7. Focus on Account Usage
Finally, a strict focus on usage prevents a value-centric approach to cloud management. “Accounts can span use cases and different priorities — yet all cost appears as a line item,” Williams said.
How to Shift to a Value-Focused Approach
Moving to a value-focused approach to FinOps is akin to the “shift left” mindset that is increasingly popular in security and other IT domains that have historically been dealt with via lagging indicators.
Some organizations try mandating discipline with regard to cloud usage — the “do this or else” approach. “That never works,” Williams said.
Instead, consumption policies and technical guardrails need to be implemented before resources are ever provisioned and deployed.
“This needs to be done while infrastructure is being orchestrated to associate them to the workload, activity and application for which they are used,” Williams said.
That’s a key use case for Torque, Quali’s cloud optimization product. Quali isn’t a FinOps company, per se, but Torque’s feature set is designed to deliver real-time visibility into cloud consumption as resources are orchestrated and deployed — a key to moving to a proactive strategy instead of doing forensics only after the invoices arrive.
Torque enables FinOps teams to see costs by the cloud provider, time used, purpose and user in real-time; understand usage cost, budgeted cost, potential savings, and actual savings; align this cost visibility with business purposes, such as a specific project, application or business process; and forecast future spending based on actual historical usage, predicted usage, usage mapped to budget trending and usage guidance.
Technology alone won’t solve the challenges many FinOps teams face today in moving to a value-centric strategy. Williams shared tips for organizations that are still in a reactive mode today:
Don’t treat FinOps like an island. If cloud cost management is done in a silo, it will be virtually impossible to develop the collaborative processes and culture you need.
Get a grip on what you’re using. It’s table stakes for a value-oriented FinOps strategy. Similarly, don’t wait for the bill to come in to identify and eliminate waste — do so in real-time, as cloud resources are being consumed. Measure unnecessary/wasted resources against business priorities, not just raw usage.
Understand why you’re using something. If you struggle to associate cloud resources with business priorities, that’s a red flag.
Don’t analyze cloud costs simply by account. Leverage technology and processes — ideally with a heavy dose of automation — that enable you to connect cloud usage with business purposes.
Implement governance and technical controls. You should be able to view and manage your resource consumption in real-time, as well as track changes and updates. Williams also stressed the need for technology tools that assume a diversity of maturity and expertise among your team. Your tools should make it simple to understand what is, ultimately, a complex environment.
And remember: Some of this takes time to get right, especially if you’ve been using a highly reactive approach — aka, waiting for the bill to arrive.
Said Williams, “You have to have a certain amount of discipline, organizational collaboration and processes established to be able to do this within a FinOps practice.”