Finops is a buzzword with many meanings, depending on whom you talk to. I’ve heard it defined as everything from “technology and tools” to “culture and communications,” and even “business processes.” It’s bigger than any of those—and sometimes part of all three.
The consensus for those dealing with cloud computing? Finops is related to cloud financial management. It’s a collection of operating models and technologies that combine best practices with technology and culture to improve an enterprise’s ability to optimize cloud costs. It’s a collaboration of IT, developers, finance, procurement, and others in the business to find the most value from cloud computing technology.
Enterprises often miss some of the larger and more valuable concepts in their finops strategies. Sometimes finops teams follow the wrong leaders, or they overthink or underthink the purpose of finops. Here are three suggestions that will help you make the most of your cloud investment:
Organizational changes. As a consultant, I rarely change the structure of organizations—that’s typically a hornet’s nest. However, when the enterprise deploys finops tracking, reporting, and controls, someone often fails to make the tough calls about who should provide the centralized control that many finops plans call for.
Setting up a cloud business office is often the best path to get to a functional finops program. It’s one thing to track spending and another to set limits and control spending using automated tools.
Someone must make those calls and then clearly communicate why the processes exist and their business value, and then keep open channels of communication with all those affected. This includes engineering, finance, and even the users and executive team. Someone or some team needs to take on this responsibility or else the finops program won’t work.
Automation. We often define finops as cost monitoring, cost controls, and cost optimization. But far too often we rely on humans to carry out the processes that we define around these concepts. That system won’t scale.
A fundamental reason to do finops in the first place is to set up automated processes that gather cloud spending data, analyze that data, and then take corrective actions. Corrective action might be spinning down compute and storage instances that are no longer being used, or sending out email warnings when public cloud spending is about to exceed the agreed-upon budget. You need to automate all of this, and it’s not that hard to do.
Visibility into cloud provider agreements. Most cloud providers make you sign agreements that specify how they bill you. Some are straightforward, such as charges for hourly usage or the amount of data used or stored. Others are confusing and not fully visible to the finops processes and tools that monitor cloud spending.
Some are so confusing that the people who set up finops processes often oversimplify the terms of the agreement. That means you’ll end up with a cloud bill that’s many times more than what you expected and/or what the finops tools forecast. This usually happens when you reach a limit or trigger some other event where the costs significantly go up or down. If you sign an agreement, you need to carefully enter the terms into your finops tool and models, including terms that are… well, creative.
Don’t overlook these three key factors that could make or break your finops system. Set up a cloud business office and assign responsibilities. Automate as much as you can. Examine and interpret the terms of your provider agreements into language the finops system can understand. Clearly communicate finops goals and objectives to the people involved, at the levels they need in order to understand the goals and objectives.
Easy, right?