Does it seem like a sizable portion of your organization’s cloud budget is disappearing into thin air? Here’s how to bring some of that money back home.
Enterprises worldwide spent a record $34.6 billion on cloud services in 2020’s second quarter, up roughly 11 percent from the previous quarter and 30 percent from the same period last year, according to research firm Canalys.
What isn’t known is how much of that money was wasted due to poor planning, faulty execution, inadequate oversight and internal conflicts, among other causes. Maintaining firm control over cloud spend is essential to curbing waste, says Jimmy Pham, a principal specializing in DevOps, platforms as a service and containers as a service at business advisory firm Booz Allen Hamilton. “Gaining control over cloud spend can’t be a one-time effort,” he warns. “A combination of technology, operations and practices can help reduce spending waste.”
Are you ready to optimize your cloud investment? Here are seven ways to keep the cloud from unnecessarily squeezing your budget.
Cloud waste hides in many forms. “There are unused resources, like storage and … other computing resources,” says Tristan Morel L’Horset, North America cloud and infrastructure growth lead at consulting firm Accenture Technology Services. Another wasteful practice, he notes, is failing to cycle application environments, including development and testing areas, when not in use. Oversizing computing resources is another effective method for squandering money, although one that can be easily fixed. “As applications move to cloud, we can now truly size for only what we need and change as demand increases,” L’Horset explains.
The overprovisioning of infrastructure services, such as leaving VMs on when not in use, or making non-production environments available while idle, is still relevant in the cloud, says John Yeoh, global vice president of research for the Cloud Security Alliance, a nonprofit that advocates cloud security best practices. Before shifting to the cloud, enterprises built and maintained such resources on premises, so the cost savings from capital and operational maintenance could be seen immediately.
“But operational efficiencies can still be optimized by utilizing the auto-scaling and the on-demand nature of cloud,” he notes.
One of the best ways to gain control over cloud spend is by closely monitoring workloads and applications and making adjustments based on both current and planned needs. “To fully optimize cloud spend, you need to have the entire picture of your organization’s cloud activity so you can make the right decisions to optimize that activity,” says Brian DeMeulle, executive director for enterprise architecture and infrastructure at the University of California at San Diego.
Automated cloud monitoring and management tools and services are essential to detect potential waste. DeMeulle says his organization uses a multi-cloud management platform and a third-party cloud financial operations organization to give IT and finance team members full visibility into the university’s cloud use, as well as to supply customized recommendations for how and when to make changes to optimize costs. “With the right combination of solutions and partner support, organizations can quickly gain control over their cloud spend,” DeMeulle promises.
Pham advises conducting regular finance and software development reviews. “This includes both recurring manual reviews and, more importantly, automated reports, alerts and triggers that are fine-tuned to provide relevant and accurate data points for quick, decisive actions,” he says.
Constant vigilance also requires using a monitoring tool that can quickly detect anomalous events. L’Horset recalls working with a client that was puzzled by inexplicably mushrooming cloud consumption.
“The infrastructure was performing very well, but one of the developers had expanded several cloud resources to perform bitcoin mining,” he explains. Only after the organization’s monitoring technology alerted IT to an anomalous consumption event was the issue was discovered and resolved. “Monitoring for consumption events has become very important in cloud management and needs to be done on a real-time basis to avoid large consumption surprises,” L’Horset says.
A consistent approach to cloud adoption generally delivers the best long-term value. Moving enterprise services to the cloud on a whim, or as a knee-jerk reaction to an external event, frequently leads to long-term waste.
The COVID-19 pandemic, for example, created a cloud-rush that some enterprises are beginning to regret. As remote work and learning environments took hold, enterprises began floating essential services into the cloud with little or no advance planning. “At first, out of necessity, it may have seemed quickest to simply move services to the public cloud or deploy new services within the public cloud,” DeMeulle says. “However, without sufficient time to plan a migration or deployment, organizations don’t think through the best consumption model.”
Effective cloud cost engineering is based on efficiency, notes Jim Haughwout, head of infrastructure and operations for music streaming service provider Spotify. “Not bureaucratic time-tracking-style efficiency, but designing systems and services to scale efficiently,” he states. Haughwout notes that Spotify was able to transform cost engineering into a platform for optimization.
Engineers love optimization, Haughwout observes. “By turning cost engineering into an optimization discipline, we tapped into this passion,” he says. “[We] gave them the insights to inspire optimizations that made their software more efficient, and as a result, faster and more scalable.”
The key to success, Haughwout says, was making cost relevant and actionable. “We provided insights that let engineers know how their services, data pipelines, websites and so on were in terms of things like cost per dollar of revenue, cost per user and cost in terms of engineer hours,” he explains. The approach enabled teams to see when their systems were performing less efficiently than desired, inspiring action and empowering the engineers to make informed time versus cost decisions.
“This converted top-down control into bottom-up optimization and planning for engineering excellence,” Haughwout says. “Once we did this, we found cost engineering started to happen automatically as part of teams’ normal work.”
To gain firm control over cloud spend, it’s helpful to begin with the final result in mind. “If you’re just focusing on cloud spend, you’ll miss out on the full value of the cloud,” L’Horset warns. Concentrate on cost, performance and innovation optimization. “A micro-focus just on cloud spend may make companies blind to the real value of cloud, which is around agility, performance and innovation,” he notes.
The expectation for any enterprise’s cloud strategy should be that it will continue to evolve and change through new landing zones, SaaS additions, fresh cloud-native services and various other developments. “To plan for something that dynamic, you need an integrated strategy that aligns cloud governance, operations and spend,” Pham says.
One of the biggest budget pitfalls IT leaders fall into is treating their cloud spend as a static item. Change happens, and cloud spend may need to suddenly grow. “This isn’t to say there should be an unlimited cloud budget,” Pham says, “just that it should be aligned into business and IT initiatives so that surges aren’t a surprise and are justifiably used to inform future spend needs and trends.”
Although cloud technology is constantly evolving, some enterprise departments may resist adapting to cloud technology needs and practices. Traditional procurement departments, for instance, sometimes have trouble shifting their practice to emerging cloud spend norms, observes Blair Hanley Frank, an analyst for technology research and advisory firm Information Services Group. “Often, these organizations are focused on driving the greatest value for money by maximizing discounts, which has led to enterprises evaluating committed spend agreements that vastly outpace their ability to consume services,” he says. “Enterprises should avoid this at all cost.”
Governance plays a particularly important role in bringing visibility and transparency to cloud resource usage, Pham notes, “because it provides clarity and accountability for cloud sub accounts and works in conjunction with ongoing reports and usage scans for sub accounts.”
It’s important to get control of the cost data, Haughwout advises. “Not just at the top-level, but down to component level,” he says. “Companies need to be able to understand what every microservice and database costs, because changes and optimizations are made at those levels.” Top-down mandates will simply result in ineffective actions that aren’t sustainable, he adds.
Look at how cloud costs figure into the business’ operating costs and set goals to continuously improve … through ongoing improvements, Haughwout suggests. Doing so will allow IT to take advantage of ongoing technological advancements while preventing teams from being slowed down in the name of “savings.”