Is your cloud spending out of control?

FinOps is defined by the FinOps Foundation (finops.org) as “an evolving cloud financial management discipline and cultural practice that enables organizations to get maximum business value by helping engineering, finance, technology and business teams to collaborate on data-driven spending decisions.” This is a relatively new practice in IT and technology…and an important one!

Cloud computing has enabled access to levels of processing power and scalability traditionally reserved for companies who field expensive data centers filled with equipment and staff. Done right, public cloud hosting allows you to quickly deploy and scale infrastructure with smaller engineering teams via automation. If done wrong, it could be an expensive proposition. Budgeting and finance planning are usually not an engineering team’s favorite topics. But one way to drive a CFO or Controller crazy is to not have a plan to manage cloud hosting costs.

The economic model for IT and tech teams has drastically changed over the last decade. Traditionally, IT and technology departments plan their budgets like any other department. After deciding on the department project roadmap for the upcoming year, budget for staffing headcount, new capital expenses (servers, facilities, etc.) that get depreciated over multiple years, and operational expenses to run daily operations (licensing, utilities, etc.) are forecast. Cloud computing adds additional complexity to this already complex equation.

The big change with cloud hosting is that your data center infrastructure no longer lands in the “CapEx” bucket as you are no longer buying equipment up front that would then depreciate over time. Instead, cloud computing treats data center hardware costs like any other utility bill (OpEx). If you run a virtual server in AWS or Azure for 24 hours, then you pay the applicable hourly rate for every hour that server runs. This presents a unique set of challenges for IT and finance teams not used to managing costs in this way.

Many organizations get burned when they start their journey into public cloud hosting. The biggest mistake is not having a plan. This could include giving the engineering teams “god mode” access to deploy whatever and whenever they want. It could be lack of cloud policies or guardrails to protect the organization. Or it could simply be that no one is on the wall watching and analyzing hosting spend.

Modern technology and IT teams love the freedom and ease at which cloud resources can be deployed and scale. No longer are we managing static “pet” servers with finite resources. Now, auto scaling ephemeral clusters and storage rule the day, removing performance barriers. Allowing engineering teams to self-service improves speed to market and project turnaround times. But one consequence of this is that without controls or guardrails, the utility-like cost of cloud hosting balloons. I have seen too many poorly engineered deployments that unnecessarily cost companies tens of thousands of dollars a month. This is usually a result of not having full awareness of how much cloud hosting costs. Engineering and product teams need to become more business savvy and ensure their architecture planning includes cost estimation.

Another mistake is simply underestimating the importance of having someone track and watch cloud spend, full time. I have paid for my salary many times over in recent years, just by devoting 20% of my focus to tracking and managing cloud costs. Not having any collaboration with your finance teams, not having proper data or analytics, and not having a plan for hosting costs are all mistakes commonly made and easily avoided.

There are several ways organizations should tackle these problems. To start, the finance and technology/IT departments need to partner up and have a strategic plan on who and how costs will be managed and forecast. Because of the auto-scaling nature of cloud resources, it can be near impossible to really say where your spending will be one year in advance during normal budget planning. This is especially true for aggressive growth companies. An analyst or manager focused on FinOps should be a full time or priority role within any organization moving to cloud hosting. This person is responsible for analytics, cost reporting, budget planning and programs around cost reduction and hygiene efforts. Ideally, this person should be technically savvy or have an engineering background with a sense for finance and seeing the “bigger picture” of business operations.

Next, analytics is the secret weapon of the FinOps role. Cloud providers, and many third parties, offer dashboarding capabilities to allow you to track spending across categories or accounts. How that is approached partly depends on the organization and how they leverage technology. In all cases, one of the best ways to track cost is through tagging. Tagging is a way to add metadata to every billable asset deployed in a public cloud provider account. Tags are virtualized asset tags. To use a Spotify analogy, you find your favorite music by searching for artist, song title, album title, etc. All those data points are metadata “tags” that you can use to filter and sort assets (your music).

DevOps or platform infrastructure teams should establish a tagging standard and procedure. Something that all teams will follow for consistency. You could add tags that reference owning department, product line, purpose, date, creator, etc. The options are endless. Once you have resources tagged appropriately, many analytics tools could take that billing data and visualize it any way that makes sense for your organization. Say you have a SaaS platform with multiple functions or services. Creating a “service line” tag would be a great way to split up costs for different components. Once you visualize how you are spending money, accountability magic can begin!

Charts and analytics tools should be leveraged to break down cost sorted based on the needs of your organization and finance department. These tools are like your monthly energy bill showing total wattage used with trends. Just like your energy bill will inform actions in your home (turning lights out at night, etc.), so will cloud analytics tools drive cost reduction projects.

Finally, holding departments or product teams to budgets can help drive spending reduction. With proper analytics and tagging policies in place, you can shine light on expensive resources or environments. This data can help drive business decisions on the viability of products and services.

FinOps isn’t a new concept, but it has become extremely important as public cloud consumption has increased exponentially. I have seen millions of dollars left on the table simply because of lack of awareness or focus in this area. Companies struggling to scale, especially in uncertain economic times, should take every step to address cost that impacts the bottom line. FinOps is low hanging fruit. Get someone on it!