During a recent webinar with the CFO Leadership Council, Matt Shover, VP of Operations & Customer Experience at ProsperOps, explained different strategies companies can use to reduce cloud costs and the best way to prioritize them for maximum “bang for your buck.” The insights are particularly valuable for teams as they begin their FinOps journey and subsequently mature through its phases.
Before we break down the top three methods to reduce cloud costs in the recommended order of implementation, here’s a quick guide for measuring success.
Start with the Right Metrics to Understand Cloud Resources and Rates
FinOps cost optimization can be thought of as spanning two universes: financial and engineering. As organizations look for ways to operate more efficiently in the cloud, they should explore efficiencies related to pricing, such as Reserved Instances and Savings Plans (financial universe) and efficiencies related to cloud infrastructure usage, such as reducing waste (engineering universe).
Before you optimize, it’s important to first get a handle on the following:
- Understanding current cloud spend
- Understanding available savings instruments
- Identifying wasteful spend
Three metrics, in particular, help you gather data relevant to how you manage cost, financial instruments, and engineering resources. In general, you want to understand:
- Cloud Spend Trending
- Effective Savings Rate
- Waste
Cloud Spend Trending: How Are You Spending in the Cloud?
This data relates to top-line spend. The goal is to understand how much you are spending in the cloud over time, break out how much you are spending by service and track it comparatively (month over month, etc.).
In this way, stakeholders have visibility into cloud costs, so that they can trend decisions over time. This is the key point: those who control the spend need to know their spend. This information is a prerequisite for accountability.
Effective Savings Rate: Measuring Your Current and Potential Savings
Discount instrument utilization, usage coverage, and various discount rates are all elements involved in calculating your savings. They are input metrics that teams are often encouraged to track, but they only tell a small part of the story. Depending on the scenario or use case, which can be very complex, it’s easy to misinterpret the data. Therefore, a useful output metric is needed.
Effective Savings Rate (ESR) is a metric that factors in utilization, coverage and discount rate to quantify savings performance. ESR is the only metric that reveals an ROI for a cloud savings strategy, like leveraging Savings Plans and Reserved Instances to reduce AWS costs.
Tracking Cloud Waste: What Can You Eliminate?
Engineering teams often use various tools and methods to identify, track, and remediate, including DIY manual approaches and recommendation-based software. Waste metrics include:
- Untagged/unknown resources
- Overutilized spend coverage
- Resources rightsize or termination opportunity
- Unused/unattached resources
With tracking mechanisms in place to understand cloud spend, savings potential, and waste, you are better informed to reduce unnecessary cloud spend.
Top 3 Cloud Cost Optimization Strategies
In most situations and for the majority of organizations, these are the top three strategies (stack-ranked) for reducing cloud costs.
1. Discount Instruments
Compute (EC2, Lambda, Fargate) comprise more than 60% of most organizations’ cloud bills. By leveraging Reserved Instances and Savings Plans, finance teams can often realistically reduce their compute costs by 30% to 40%, reducing their overall AWS bill by as much as 25%. These compute discount instruments require organizations to commit to a minimum amount of hourly usage return for a lower rate, so it’s important neither to over-commit nor under-commit. Both scenarios lead to suboptimal savings.
2. Enterprise Discount Program and Private Rate Cards
Enterprise Discount Programs (EDP) and private rate cards allow large organizations to get higher discounts in exchange for spend-over-time commitments. Many organizations will aggressively commit to these discounts—often to save a marginal 1% or 2% of their AWS bill, which then limits their ability for further optimization (e.g. discount instruments, eliminating waste, or engineering changes). Committing to a more achievable amount reduces this risk.
3. Eliminating Waste
Finally, engineering teams can also reduce cloud costs by reducing unnecessary cloud spend. These engineering optimizations take time and effort to execute, and they may involve re-architecting services, selecting more optimal cloud services, right-sizing resources, and more.
These techniques often involve a cost-benefit analysis to determine whether it’s worth the engineering effort required to implement a change. However, taking an initial pass at addressing some relatively easy wins is often effective for most organizations.
Watch our webinar replay of What, Why, and How of Cloud FinOps to gain a better understanding of these cloud cost optimization strategies and how to implement them within your organization.
How Should Organizations Implement Cloud Cost Optimization Strategies?
While there are many different approaches to cloud cost optimization, most organizations should consider using all strategies available to them, starting with those that have the greatest FinOps ROI.
Since most organizations spend 60% or more on cloud compute services, Shover recommends starting with compute discount instruments. They are generally the easiest cloud cost optimizations to implement and produce a substantial, immediate impact on cloud spend.
This is especially the case when organizations adopt an autonomous cloud cost optimization solution like ProsperOps that manages Convertible Reserved Instances, Standard Reserved Instances, and Savings Plans without any ongoing effort.
Large organizations with a very high cloud spend should consider EDPs and other discounts and credits that AWS provides. Shover emphasized the importance of negotiating EDPs: certain elements may be fixed, but others are indeed negotiable.
Finally, engineering teams should be judicious with efforts to eliminate cloud waste, noting this can take time and resources to implement. That said, cloud surveys have estimated that up to 30% of the average organization’s cloud spend is waste. While not all cloud waste can be easily addressed, Shover says most organizations have low-hanging fruit they can resolve immediately and should start by eliminating unused resources.
Start Your FinOps Journey with ProsperOps
ProsperOps is an autonomous FinOps solution that manages cloud compute discount instruments to maximize savings while minimizing commitment risk. By ingesting telemetry from your cloud environment and analyzing it in real-time, our AIOps platform can identify savings opportunities and execute them instantly, automatically returning money to your AWS cloud budget.
This hands-free approach reduces the burden on finance and engineering teams as they adopt cloud FinOps best practices. Organizations interested in understanding their cloud savings potential can request a free compute savings analysis.
