4 steps to improve FinOps and cloud cost optimization
What’s the optimal way to manage cloud finances, one of the core disciplines of a FinOps practice?
The answer largely relies on automating efforts, but many FinOps practitioners haven’t taken full advantage of opportunities available to them or of the savings available to their organizations. According to a report from the FinOps Foundation, a Linux Foundation non-profit trade association focused on codifying and promoting cloud financial management best practices and standards, nearly half (49 percent) of the more than 800 respondents (with a collective $30+ billion in annual cloud spend) had little or no automation of cloud spend management.
Of those with some automation, nearly a third rely only on automated notifications (31 percent) and tagging hygiene (29 percent); only 13 percent automated rightsizing, with fewer (9 percent) automating spot use. Similarly, the Flexera 2022 State of the Cloud report found that, while cloud spending continues to accelerate, controlling costs remains difficult. Organizations estimate 32 percent of cloud spend is wasted, which is a major reason why 59 percent of organizations plan to optimize existing use of cloud, making it the top initiative for the sixth year in a row.
Clearly, companies are missing opportunities to optimize cloud costs. Many enterprises are responding by building out sophisticated FinOps teams. For those teams to succeed, their processes need to be well defined. Here are four ways to gain control of your cloud financial management initiatives and build out an efficient FinOps practice.
- Take a total cost of ownership view.
Companies use many different clouds, with many relying on both public and private clouds. A single organization might use AWS for one workload that leverages Lambda, Google Cloud for another that leverages machine learning (ML), and Alibaba for a workload in China, for example. Whatever multi-cloud configuration is in use, one of the results may be significant challenges in evaluating expenses. Taking a total cost of ownership (TCO) view helps ensure that you’re seeing, managing, and finding efficiencies in the big picture.
Gaining a comprehensive view of costs requires having a common bill ingestion interface that can show and aggregate all costs, from public and private clouds, including on-premises and hybrid cloud spend. When using a common bill ingestion interface, you can specify (in CSV format, for example) any costs to be ingested. These may include costs for a content delivery network (CDN), hosted services, or for services from across clouds (including costs from different regions). Ingestion of other costs, such as labor costs and on-premises charges (e.g., licensing) or container costs (e.g., KubeCost), augments your ability to gain a robusts and clear TCO view.
- Prepare for effective multi-cloud reporting.
The natural next step after ingestion is reporting. Multi-cloud reporting relies on normalizing your cloud spend. Normalizing spend may mean eliminating certain costs (like AWS tax), marking up or down your bills to reflect internal adjustments, and/or currency conversion so that all of your spend is in a single currency (with exchange rates adjusted monthly). Once this spend is normalized across all of the vendors you use, you’re able to generate insightful reports.
Effective multi-cloud reporting allows you to drill into data to see, for example, the different cloud vendor services. This is another important area where normalization is required. Vendors name their services in very different ways, complicating the ability to gain clarity into cloud costs. Take the step of recategorizing these services for consistency across cloud service providers. Doing so allows you to take a uniform view of compute, network, and storage services across the different vendors. This can help in your initiatives to build customer reports that meet your showback and chargeback requirements.
- Get the right data to the right people.
An effective FinOps program relies not just on ingesting data from API-enabled sources -- cloud vendors or external hosted solutions, such as a configuration management database (CMDB) or an IT service management (ITSM) tool, for example. What’s necessary is the ability to analyze and manipulate that data in a manner that supports taking action.
Applying automated policies can help make this process as streamlined as possible. Rely on policies -- based on best practices and customized to meet your organization’s specific needs. These may address categories such as cost management, compliance, operational security, or SaaS management; other policies may be specific to cloud vendors.
When creating policies, identify what credentials are required, how frequently the policy should be scheduled (e.g., every 15 minutes, weekly, monthly), and define relevant credentials. Possible actions to take with the output of these policies include anything that can be carried out via an API. This might mean simply raising a ticket for engineering, producing an email, or requesting an approval. Other common actions with a cloud vendor may be to tag or label workloads, terminate instances, or rightsize infrastructure. Each of these actions is a step toward streamlining your cloud cost optimization efforts. Integrating these policies with Jira or ServiceNow, for example, helps ensure that the output of your policies gets to the right people within the organization, helping them take action on them.
- Define and track your key performance indicators.
Clearly necessary, but often overlooked, are key performance indicators (KPIs). Take the time to define and track the KPIs that will explain the success (or shortcomings) of your optimization efforts. Common KPIs for FinOps include unit economics, which may include cost per instance hour, the total number of transitions, or total bill per vendor. KPIs are useful in how they not only measure outcomes, but detect outliers in your data. There’s a story in your data; make sure you’re paying attention to it. When you encounter results that don’t benefit your bottom line, rethink your approach and remediate accordingly.
Managing technology spend is tricky enough. There’s no need to complicate the cloud cost optimization process itself.
Jeremy Chaplin, senior cloud solution architect at Flexera, specializes in cloud financial management & solution architecture. Before joining Flexera, he was cloud solution architect for CloudHealth by VMware. Jeremy is a member of the FinOps Foundation.