When organizations begin making the transition from on-premises IT infrastructure to public cloud, their financial IT operations change as fundamentally as their infrastructure does.
The largest conceptual paradigm shift is that from large, infrequent capital outlays for IT infrastructure procurement to hourly operational expenses. In the former case, most organizations have a structured procurement process that requires various levels of approval before outlays can take place. By contrast, public cloud services are procured in real-time by individual technical teams within the organization, de facto eliminating much of the oversight provided by procurement procedures. Procurement de facto becomes federated to the consumers of the services.
In many organizations, these teams may be tasked with conforming to a budget, but services can and often are purchased in excess of the budgeted amounts. These overruns are incurred before remediation can take place — the direct opposite of what occurs in a capital expenditure environment with a structured procurement process. This federated procurement process and the natural complexity of public cloud billing can combine to pose significant challenges for organizations that have yet to establish a disciplined FinOps team. Here are some of the most common challenges we see.
1: Inability to Trace Cloud Costs Back to Business Units or Other Logical Financial Categories
The process of tracing costs to business units, products, or customers is often referred to as “chargeback.” The key to effective chargeback lies in applying the correct structure of billing accounts, folders, projects or labels to the cloud infrastructure. When these structures fail to line up with business units, it can be extremely difficult to accurately report financial results or to make business decisions based on the cloud cost data.
When a designated FinOps team has yet to be established, no one is formally accountable for seeking solutions to remediate the billing data and/or conduct post-processing in an attempt to meet management’s reporting needs. The severity of these challenges ranges from moderate to severe. Prior to engaging with us, up to 20% of cloud spend could not be traced to any business unit in one of our large Fortune 100 clients. In that case, the 80% that proved traceable could only be charged back after an enormous effort was exerted each month in post-processing of the cost data — post-processing that was so complex that much of it was outsourced to a managed service provider. Despite the scale of the effort exerted in the process, the resulting financial reports proved to be error-prone and were deeply distrusted by finance and management.
When a designated FinOps team has yet to be established, no one is formally accountable for seeking solutions to remediate the billing data and/or conduct post-processing in an attempt to meet management’s reporting needs.
An inability to segment and charge back cloud cost data also destroys an organization’s ability to make good business decisions using the cloud billing data. Consider, for example, trying to make product development decisions without being able to measure the public cloud component of a product or service’s cost. How can decisions be made about pricing products or services if the public cloud component of a customer’s cost can’t be measured? How can individual deals be quoted accurately?
2: Inefficiency, Waste, or Over-Provisioning and Resultant “Crisis Mode” Optimization Initiatives
We were on a consultation call with a large, prestigious management consulting firm and we asked how they charged back their cloud costs without a FinOps team. Their answer? They didn’t. Their large public cloud spend all simply went into a single bucket: IT Infrastructure. Next, we asked what challenges they encountered with over-provisioning or waste. Their answer: they didn’t perceive any particular challenges. They had no reason to suspect that they weren’t operating efficiently. Since they didn’t perceive any particular challenges, they didn’t see a particularly strong need to stand up a FinOps team.
They were almost certainly wrong about how things were going. A public cloud consumer that 1) doesn’t segment their cloud spend and 2) doesn’t perceive any particular problems with waste or inefficiency can be compared to a driver speeding down a road at night with no headlights on. The driver can’t see anything, so everything must be fine!
Only unlike the driver, the cloud consumer has no idea how much damage they are doing. We have watched many times what happens when those “headlights” get turned on for the first time. When clients initially deploy a cost management tool such as Ternary or Cloudability they often see an immediate 10-15% drop in cloud spend before they even stand up a FinOps team or embark on a single optimization initiative.
When it happens, here is the cause: individual teams within the organization are given logins to the cost management tooling that provides a view of the costs they are individually generating. When they are given this visibility for the first time they instantly see where they are wasting and begin shutting down or deleting unnecessary resources. From there a FinOps team is established and more concentrated optimization initiatives begin — often resulting in a doubling of the savings already realized. We have seen it time and time again.
So the axiom works like this: when cloud consumers don’t have visibility into their spending, there is a near certainty that they are wasting money, because the technical teams generating the spending can’t see what they are doing. And the single most important component of measuring efficiency is segmenting it in a way that 100% (ideally) of the cloud spend can be traced to humans who are responsible for having generated it.
3: The Crisis-Mode Roller Coaster
Usually in the scenario where cloud spend is not segmented the first in a series of internal crises will occur when the cloud costs get large enough to draw the attention of senior leadership, who concludes (correctly) that the bills are unnecessarily high. A mad scramble then ensues, with groups of engineers individually trying to develop up their own optimization methods. Things often calm down when they find enough savings for senior management’s attention to shift elsewhere.
Unless the organization establishes a FinOps team, however, this crisis will soon repeat when discipline again lapses and costs begin to drift back upward. FinOps teams are often established when organizations find the series of repeated crises has become so exhausting that they realize they need to adopt a new, formalized approach to the problem.
4: A Lack of Governance and Knowledge Sharing
Without a well-trained and disciplined FinOps team, who is responsible for generating guidelines and guardrails for use of public cloud services? Well-run cloud deployment will have quotas to prevent inadvertent spending blooms, for example, and anomaly alerting to provide early warning if and when any such blooms occur. Labeling taxonomies and/or project naming conventions are established to facilitate financial reporting and these taxonomies are evangelized to the technical teams to maximize labeling coverage. Technical teams are educated on the basics of cost optimization and KPIs are established for things like Reserved Instance/Savings Plan/Committed Use Discount coverage. Without a FinOps team to pursue each of these objectives, best practice disciplines will only be sporadically applied, and the enterprise will encounter a predictable range of negative consequences.
Do These Challenges Sound Familiar? If So, Here’s What to Do Next
- Nominate candidates from technical teams, finance and accounting to dedicate part of each week to FinOps activities. The best candidates already have some exposure to dealing with cloud billing and so have a solid introduction to the challenges and effective solutions. If your spend is large enough, consider making one or more full-time FinOps hires or re-purposing an existing team member full-time to the function.
- Take advantage of the FinOps foundation! The Foundation has posted a wealth of free FinOps resources on their website, offers regular knowledge-sharing summits for members, and offers terrific self-paced or live training. They are your best resource to train up your FinOps practitioners and keep them abreast of the most recent best practices.
- Kickoff FinOps initiatives focused on optimizing each of the following categories:
- Visibility and reporting
- Operating efficiency
- Pricing efficiency
- Budgeting and forecasting
- Remember that the best FinOps organizations provide resources to help the technical and finance teams reach their goals. Although governance and accountability are key to the best FinOps practices, it is best to avoid the appearance that the FinOps team is intended to be an “enforcement” body. Rather, the FinOps team is there to arm the technical team with the knowledge and resources they need to establish their own governance practices.