re:Invent 2022: FinOps Metrics That Matter

FinOps metrics help surface waste, show when rising cloud costs correlate to innovation and profitability, and frame spending in business value.

At re:Invent 2022, Apptio hosted the speaking session “Metrics That Matter: Navigating Financial Uncertainty with Actionable Cloud KPIs”. Ashley Hromatko, director of business operations at FinOps Foundation; and Vik Saluja, cloud strategist and FinOps manager at Cigna; joined Marisa Banigan, sr. director of product marketing at Apptio to discuss the importance of using FinOps metrics to measure cloud cost efficiency and guide decision-making, particularly in times of increased budget scrutiny.

Around the globe, organizations are navigating a period of financial uncertainty. Coming out of the pandemic, we saw inflation hit a 40-year high, persistent supply chain issues, and geopolitical conflict, all project us into an “economic limbo.” But we shouldn’t forget that uncertain moments can inspire a great deal of innovation. In fact, during the 2008 global recession, Technology Business Management was formalized as a practice to help IT leaders navigate budget swings while still fueling growth and innovation. That balance is still critical today, as market leaders do not simply rely on innovation to sustain their success. Instead, they genuinely drive competitive advantage when they find intelligent ways to reduce costs, maintain margins, and focus on their customers in parallel.

Cloud is your secret weapon to beat your competition

Public cloud services provide a unique opportunity to increase cost efficiency and strengthen your organization’s competitive advantage. This is because every minute in the cloud matters from a cost perspective. Costs associated with cloud are variable operating expenses: it’s far easier to impact these types of costs than capital expenditures or fixed costs — and changes made will have an immediate impact. When your team turns off or rightsizes virtual machines (VMs), you will realize cost benefits in the seconds or minutes that follow — cost benefits in seconds or minutes are essentially unheard of throughout the rest of your technology investment portfolio!

We also benefit in the future from making smart efficiency improvements or optimizations. The waste you remove today is waste (and associated overspend) that no longer exists tomorrow and into the months ahead.

Cloud provides us with levers in both usage and pricing. We should evaluate both our cloud procurement and optimization decisions and find opportunities both for waste reduction by deprovisioning underutilized instances (that’s the usage impact) but also find opportunities for pricing efficiencies gained from savings instruments like Reserved Instances and Savings Plans, which reduce hourly rates without the need of engineering resources.

Finally, reducing cloud costs comes with fewer downsides, like less organizational disruption and less impact on company culture and innovation, than traditional cost reduction actions like hiring freezes, budget slashes, or, unfortunately, as we are seeing today — layoffs. These actions impact your organization in the moment and can have a far longer recovery time.

Guide your cloud program with FinOps metrics

FinOps metrics can help measure and drive efficiencies within your cloud program, helping your team to avoid overreacting to budget scrutiny, stop wasting time on low-impact changes, and, most importantly, avoid suppressing business growth in moments of financial uncertainty. By measuring cost efficiency, cloud program TCO, and unit cost, cloud practitioners can lead with certainty and exploit the cloud as their organization’s secret weapon to cost reduction without sacrificing long-term growth.

Metric #1: Public cloud cost efficiency and optimization

This metric surfaces indicators of cloud waste. The fact that every minute in the cloud matters is a double-edged sword. Savings are immediate, but so is waste and associated overspend. You may overspend when you provision a resource, but the meter also runs while you look for and turn off resources. Efficiency KPIs can help show you where wasteful spending is happening and spur rightsizing action.

Think about setting optimization metrics in two categories. First, look at the rates you are paying the cloud vendors; second, look at your cloud usage. Commitment programs give you the best rates from your cloud providers. But how can you tell if you are maximizing the value from these instruments? You know when to take action when you have clear metrics to measure both the current commitment coverage and utilization. Many Apptio customers will set a target on coverage and utilization to ensure they are getting the most of their cloud spend and maximizing the savings they can achieve through RIs and Savings Plans. For coverage, 70%-80% is a great coverage rate — meaning 70%-80% of your applicable usage is covered by RIs/SPs, with minimal running on demand. For utilization — getting as close to 100% as possible is usually what our customers will try to target since anything less means some of the purchased RIs/SPs are going unused.

Metrics that focus on waste and rightsizing ROI drive more efficient usage. Often, a rightsizing action can seem insignificant — a $5 saving per day or instance may not seem newsworthy. But a roll-up of all realized savings gets people’s attention. Apptio has created integrations with engineering team tools like Jira to help cloud practitioners articulate the total savings realized to executive stakeholders and engineers.

During “Metrics that Matter: Navigating Financial Uncertainty with Actionable Cloud KPIs,” Ashley Hromatko outlined the core principles of metric-driven cost optimization. “Metrics without targets are just pretty graphs,” said Hromatko. Automated measurement helps scale out the understanding of cloud data. “You need a proper understanding of data to determine realistic goals.”

Metric #2: Cloud program TCO

Customers’ monthly invoice from each major cloud vendor represents only part of the cost story. The most common indirect costs can include the following:

  • Fees for observability solutions help ensure operational and security integrity across a fleet of cloud VMs
  • Specialized third-party cloud technologies or tools, including content delivery networks (CDNs), serverless data warehouses, or FinOps tools like Cloudability
  • Supporting labor and additional cloud vendors depending on your cloud footprint and how you allocate internal resources

To measure cloud program TCO, cloud leaders need to bring direct and indirect or shared cloud costs together and then allocate back to the business based on team, product, application, etc. This measurement helps cloud practitioners and cost owners understand if rising cloud costs correlate to accelerating innovation and profitability or simply inefficiency.

Some of the best examples, we’ve seen across our customer base, of the benefit of bringing visibility and accountability to indirect costs relate to optimizing observability solutions such as New Relic and Datadog. These best-in-class monitoring solutions are an important part of running virtual machines, but they come with a real cost. Engineers in these teams can decide where it makes sense to run these monitoring tools (e.g., production workloads) and where it’s overkill (e.g., testing and development workloads). Reducing the use of advanced monitoring leads to real cost savings for the business.

During “Metrics that Matter: Navigating Financial Uncertainty with Actionable Cloud KPIs,” Vik Saluja described the parts of Cigna’s cloud TCO. It includes direct costs that go into cloud services for individual accounts and shared costs that go into managed services that span multiple accounts (e.g., security, enterprise support, transit gateways, and containers).

Metric #3: Unit cost

This metric looks at cloud costs in the context of the specific business value it delivers. All stakeholders understand unit cost metrics because they are metrics written in plain English (e.g., cloud cost per each dollar of sales, each customer transaction). It builds on the cloud program TCO metric and extends into costing by unit, unit economics, and even profitability or margin analysis.

An Apptio example illustrates the value of a unit cost metric. Processing billing lines from cloud bills is integral to Cloudability — and supporting some of the largest public cloud users in the world, we process a lot of billing lines! Therefore, measuring the “cost per billing line” is critical for our pipeline team to ensure we are operating efficiently. Over the past year, we have seen the number of total billing lines surge to over double the number processed last year, but through optimization efforts, we’ve reduced our total processing costs (“cost per billing line”) by 75%. Increases in cloud spend will always raise questions, but a fall in unit costs tells leadership that cloud teams are optimizing spend effectively.

Public cloud services provide a pathway to reducing operational expenses and improving efficiency. But cloud adoption must be twinned with FinOps practices — and metrics that matter — to ensure resources are retained for innovative and business-critical projects while eliminating consumption that isn’t driving business value.

Watch “Metrics that Matter: Navigating Financial Uncertainty with Actionable Cloud KPIs” to get more insights from Ashley and Vik. Also, check out our breakout session at re:Invent 2021 “Developing Your FinOps Practice to Make the Most of Your AWS Cloud Spend.