5 ways to maximize your cloud investment

Migrating infrastructure and applications to the cloud is never straightforward, and managing ongoing costs can be equally complicated. “Overspending is easy to do,” says Chris McMasters, CIO at the City of Corona, California. No IT organization wants to get caught short on processing or storage resources that could negatively affect operations, or have to suddenly add resources that exceed the budget. “You worry you don’t have enough capacity, so you overprovision,” he says.

Optimizing cloud investments requires close collaboration with the rest of the business to understand current and future needs, building effective FinOps teams, partnering with providers, and ongoing monitoring of key performance metrics. “FinOps is part of the equation, but from a CIO perspective, you need a top-down view that starts with the strategy before you talk about the components of it,” McMasters says. Plus, you need to balance the FinOps team’s need for autonomy against the CIO’s need for centralized control to gain economies of scale and avoid runaway costs.

Plan for cost savings up front

Cost optimization starts with defining desired business outcomes and architecting a cost-effective solution right from the start, says Jevin Jensen, research VP, Intelligent CloudOps Market at IDC. Refactoring applications to take advantage of cloud-native services is vital to maximizing cloud ROI. “Refactor your applications to take advantage of web services or serverless capabilities, and re-architect your infrastructure to optimize resource usage,” he says. In this way, you can take advantage of the cloud’s agile, on-demand approach with unlimited capacity without breaking the budget.

Be strategic with reserve pricing

Reserve pricing for cloud services can reduce resource usage costs by as much as 70%, says David McKee, who in his role as fractional CTO, tech founder, and digital twins thought leader at Counterpoint Technologies acts as a part-time CTO for nine companies in the US and Europe. “They’re giving you a discount because you’re committing to a fixed amount.”

But many CIOs, worried about going over budget, pre-book too much capacity. For example, organizations may commit to containers that are too large, or reserve too much disk space. While having that cushion avoids unplanned budget issues down the road, many CIOs waste money on substantial amounts of pre-booked capacity they never use, McKee says. Usage estimates need to be more accurate, and cushions should be smaller, he says. “Yes, you pay a premium if you go over, but in many cases, even if you do go over a bit, you’ll still save money.” Understanding the price differentials between reserve and regular pricing is key to performing that analysis.

Over the years, McMasters bought overcapacity and hoped he had enough. Today, he says, they operate more accurately. To gain that insight he monitors ongoing usage and meets weekly or biweekly with internal business leaders — and with Microsoft, the city’s primary cloud service provider, to review current and future needs.

“Understand your licensing schemes and usage policies, which can be very complicated,” says McMasters. “What exactly happens if you go over and what will they charge you?” But while McMasters’ forecasts are now more accurate, he still likes to have a cushion: he gives himself a 10 to 20% margin when pre-booking capacity, he says.

Measure often, monitor constantly

Proper insights require both a knowledge of desired business outcomes overall and for each business unit, and ongoing monitoring of key metrics. McMasters takes full advantage of the analytical tools Microsoft provides with Azure Cloud Computing Services to assist with cost optimization analysis, but he doesn’t stop there.

“We also use Microsoft Power BI, a customer experience platform to survey customers, and we review support tickets to determine where any problems may lie,” he says. “We need hard metrics because we’re running 800 instances of cloud computers. It was difficult at the beginning because there are many capacity usage unknowns, especially when you’re dealing with hundreds or thousands of servers. Our biggest mistakes were overprovisioning or thinking we had enough CPU, and having an application grind to a halt. It took a few years to figure out what our capacity needs looked like.”

Automation is also paramount here, says the CIO at a major Europe-based manufacturer who asked that his name and company not be used. “We implement automation solutions to generate recommendations to the development teams, who are our users, and we implement savings plans for most accounts — especially for the hyperscalers,” he says. Then there’s housekeeping. His team conducts annual campaigns to contain costs by cleaning up unused accounts, reducing space requirements, and releasing unused components, for example. “We also review SLA with our end users and service providers, and implement policies and other customer measures based on license agreements and the capabilities of the providers,” he says.

Another way to save money: allow infrastructure to wind down or operate at reduced capacity during off hours, says McKee. “Most organizations reject this, but practically speaking they can do it,” he says. For example, retail sites typically have times when very few people shop.

As for metrics, the European manufacturing company has a centralized monitoring team for all cloud services that measure general specifications, such as capacity, number of users, subscriptions to services, and availability. “But there are also vendor-specific metrics we define, and we build telemetry using tools based on usage and needs,” the CIO says. “For example, if a cloud vendor hosts a data lake that requires operational technology data to synchronize and feed back into a decision algorithm on the production line, we measure latency. But for other tools where latency isn’t critical, we don’t measure it.”

Cultivate relationships with internal business leaders

“Understanding the business value brought by the cloud, and its impact on the current and future status of the business, is crucial to defining long-term strategy at the lowest price point,” says Radu Vunvulea, group head of cloud at London-based software company Endava, adding it also ensures that licensing aligns with the business model and lifecycle of the associated systems. Achieving that alignment requires a high level of communication and collaboration across all departments.

For CIOs, that means building strong relationships with the C-suite and business unit managers so they can plan for and optimize cloud spend, says McMasters. “What’s the business case for use of the technology, and the strategy for a two- to three-year period, and where do we need to be two to three years from now? That helps me plan.”

FinOps teams are critical to success, says Vunvulea. “They provide CIOs with a transparent overview of cloud spend across service providers, teams and projects,” he says. “Having a dedicated team responsible for tracking and optimizing cloud spend allows CIOs to understand how cost deviations align with overall business and revenue objectives.”

Jensen says many large organizations already have built these organizations. “In this quarter, 74% of large enterprises report having FinOps teams and processes in place, up from 61% in 2022,” he says. Those enterprises reported a 30 to 35% cost savings in the first year after implementing FinOps teams and tools.

“Companies often put FinOps teams in place to address cost and unsatisfactory cloud ROI issues after they arise,” says Jensen, but you need to have those teams in place up front, before making new investments. “Failure to collaborate and hold cross-functional teams accountable for expected returns means costs can spiral out of control.”

“Keeping cloud costs solely in IT is a recipe for out-of-control spending because the business reaps the benefits while IT shoulders the expenses, and the blame, when they go over budget,” Jensen adds. “The best arrangement is to have a full chargeback of cloud costs to the business unit or group requesting those resources.”

How you organize FinOps teams can greatly affect cloud operating costs, says the Europe-based CIO. “Whether agile should include FinOps or not is a matter of debate,” he says. He learned from experience that you can’t let every agile team choose its own cloud providers, select its own services, and place its own orders. “If you go all the way and implement FinOps in an environment where you have platforms used by more than one team, you lose economies of scale. This creates a lot of missed opportunities, and is one of the debating points with agile: without end-to-end oversight on certain aspects, how can you reach economies of scale? You also sacrifice the opportunity to standardize on optimization, operations, security management and so on,” he adds.

With this in mind, he and his team consolidated all platform operations, including financials, used by more than one team. So they created an XaaS team, called the Anything as a Service team that owns cloud operations best practices, overall account management, and the financials. The XaaS team still lets individual groups do their own thing, although it does give them the option to have XaaS manage the service. “But if the service or vendor they want to use is used by other groups, then there’s no choice,” he says. Ultimately, the platform team owns the financials as well as platform operations, such as patching, security and optimization. The requesting team works only at the application layer. “That’s the compromise we’ve built to respond to the conundrum between DevSecFinOps versus economies of scale,” he says.

Finally, watch out for resistance within teams when it comes to optimizing cost efficiencies. “There’s often a defensive view toward being efficient with our infrastructure spending,” McKee says, and that attitude can quickly spread across the organization. The solution lies in making those teams understand they’re getting involved in innovation, not just a cost-cutting exercise.

Get personal with your vendors

Typically, only companies with multi-year commitments and savings plans get the personal attention of hyperscalers, says Jensen. Having such commitments in place increases cost effectiveness because personalized feedback on proper resource configurations and new web services can improve application performance and reduce costs. “FinOps teams should meet monthly to discuss recommendations from their hyperscalers,” he says.

Open and transparent communication with vendors is essential. It’s important to remember these organizations are run by people, so building relationships at the right levels in each organization is key, as well as holding regular meetings with sufficient and transparent data from both sides to ensure mutual benefits with no resentments. McMasters, for instance, keeps up with each cloud vendor’s roadmap, and looks for opportunities to collaborate. As an example, Microsoft wanted to make inroads with municipalities for its virtual desktop infrastructure, so it was able to work out a win-win arrangement to provide those services. “Know what’s the bigger win for them and what’s on their roadmap,” he advises. “Are there products or things on their roadmap that benefit both parties? In this case they were looking for opportunities to prove themselves.”

He also reviews the overall health and direction of the city’s cloud service providers, which is particularly important for smaller providers and startups. “We look for transparency,” he says. “Providers that are very transparent tend to have a solid roadmap and we work best with them.”

And when reviewing contracts for renewal, McMasters recommends hiring an analyst firm to help review changes to licensing and other cost factors. “Having a third party helps you level the playing field,” he says.

The technical aspects of cloud computing are the easier part when it comes to cost optimization—and the part many CIOs feel most comfortable—but cultivating relationships with vendors and especially internal constituents, is the ultimate key to success.

“To me, the biggest part of cloud optimization is understanding your own organization,” McMasters says. “It’s easy to misstep and overcompensate, especially on the tech side.” Optimizing cloud spend requires getting the most from the technology and building strong partnerships with your vendors. “The best relationships we have aren’t with the tech providers, however, but with our other executives,” he adds. Only when that’s in place can you anticipate needs and provision accordingly.

Establishing cloud governance and FinOps teams is also critical to get the information CIOs need to ensure cloud operational costs continue to align with business imperatives. Those teams should measure the efficiency of cloud spend, optimize costs, analyze spend, and issue reports to leadership, says Vunvulea. CIOs should then use that information to measure cloud value relative to business operations. “Understand how business KPIs map back to the cloud, and how cloud value should be measured from the business point of view,” he says.

To achieve that, FinOps teams should leverage available toolkits for cost analysis and optimization, conduct annual clean ups to eliminate unnecessary cloud resources, and monitor progress, says the European CIO.

“Discover where you’re currently spending to make sure you can explain all line items, improve spending where immediate action can be taken, and monitor for continued improvement,” says McKee. And make sure your teams see cloud cost optimization as a way to innovate, not just a cost-cutting exercise.