How to bring runaway cloud costs under control

It’s no secret that organizations are deploying more and more cloud services as they look for ways to increase business agility and scalability while reducing costs. But in an interesting twist, increased cloud adoption can lead to runaway spending. As a result, enterprises need to look for ways to control their cloud costs.

A May 2024 report by research firm Gartner said worldwide end-user spending on public cloud services was forecast to grow 20 percent and total $675.4 billion in 2024, up from $561 billion in 2023. The market growth is being driven by the rise in generative AI and application modernization.

With continued growth, Gartner expects public cloud spending to eclipse the trillion-dollar mark before the end of this decade. All segments of the cloud market are expected to see growth in 2024, the firm said.

While a rise in overall spending on cloud services doesn’t necessarily translate to higher cloud costs at individual organizations, it’s safe to say that many enterprises are seeing a cost increase. In some cases, there might even be out-of-control expenditures that cut into the financial benefits of the cloud.

7 ways to control cloud costs

Experts offered the following suggestions for reining in cloud costs:

Ensure cloud services match requirements

Take advantage of reserved instances

Leverage analytical tools to monitor usage

Use resource tagging

Closely examine billing documents

Deploy storage autoscaling

Create a culture that embraces cost control

Ensure cloud services match current requirements

Do all the cloud services an organization has in place meet its current needs? Not likely. When companies procure cloud services, it’s generally to meet a specific need or to take advantage of a market opportunity. Sometimes, this is a short-term requirement.

For example, a business might need to launch a big data analytics project before introducing a new product or service in the market. A cloud-based service provides the means for collecting and analyzing customer and market data, providing the necessary computing resources.

Once the project is completed, those cloud services are no longer needed, however.

“A critical practice we implement is ‘right-sizing,’” says Simon Bacher, CEO and co-founder of technology startup Ling, which relies heavily on cloud services to run its business.

“We review our cloud usage regularly, making sure our chosen services match our unique requirements,” Bacher says. “When we launched Ling, we started with a basic package, and as we grew, we scaled up appropriately. This prevented us from over-spending on services we didn’t require in our early stages.”

A big part of ensuring that cloud services match up with organizational needs is to review cloud service usage in an ongoing way.

“We exercise caution to avoid overcommitting and periodically reevaluate [cloud spending] as usage changes,” says Markus Schaal managing director and COO at software company Plitch. “To further reduce waste, we right-size instances and continuously track resource consumption across environments.”

Take advantage of reserved instances

Amazon Web Services (AWS), one of the big three public cloud providers along with Microsoft and Google, offers Amazon EC2 Reserved Instances (RI), which can provide a discount on services up to 72% compared with on-demand pricing, according to AWS. It can also provide a capacity reservation when used in a specific availability zone.

EC2 RIs provide a discounted hourly rate and an optional capacity reservation for EC2 instances, the company says. AWS billing automatically applies a customer’s RI’s discounted rate when attributes of EC2 instance usage match attributes of an active RI.

Microsoft’s Azure and Google Cloud also offer their own versions of reserved instances for customers.

Ling takes advantage of this arrangement to save on costs. “After evaluating our long-term needs, we committed to a one-year plan for some of our primary cloud services, taking advantage of the significant discounts that come with reserved instances,” Bacher says. This saved the company about 15 percent annually in costs.

Another company, Internet proxy services provider Live Proxies, upon reviewing its cloud usage patterns found several workloads that would benefit under reserved instance pricing.

“By committing to one-year or three-year reserved instances, we could bring down our cloud spending by up to 40 percent,” says Jacob Kalvo, co-founder and CEO. “This is a solution that involves some sort of up-front commitment, but pays off in the long term with considerable savings. For example, our data scraping operations, running continuously, were perfect candidates for reserved instances. This results in much more predictable and manageable cloud budgets.”

Leverage analytical tools to monitor usage

“One of the most potent strategies for managing costs in the cloud is to put robust monitoring and analytical tools in place,” Kalvo says.

“These are mechanisms we put in place to get very sophisticated details on how we utilize the cloud,” Kalvo says. “This helps us discover underutilized resources and removes waste. Continual vigilance over your cloud environment keeps you on a trajectory toward cost reduction via data-driven decisions.”

For instance, Live Proxies discovered that some of the virtual machines (VM) it had in use were running continuously, even though they were only needed during business hours. Scheduling these VMs to shut off outside of core hours yielded significant savings, Kalvo says.

“Gaining visibility into expenditures is very important,” says Puneet Gogia, data and business analyst and founder of Excel Champs, which provides Excel certification. “Without the proper tools, it’s easy for costs to spiral out of control without noticing.”

Proactive tools and strategies for managing cloud spending “ensure you’re using resources effectively without waste,” Gogia says. “The returns come from digging into patterns, committing to cheaper options, and keeping deployments tuned.”

One client of Excel Champs was overprovisioning resources and spending more than needed without realizing it. By implementing tools available from AWS and Azure, it was able to see where spending was going and this led to a 20% reduction by cutting waste, Gogia says.

Use resource tagging

Cloud resource tagging, a process that involves attaching descriptive metadata or labels to cloud resources to define the types of resources within a cloud environment, can be an effective way to help optimize cloud costs.

“Resources tagging gives visibility on how expenditure in the cloud can be traced or managed by the various departments, projects, or teams,” Kalvo says. “We have a very strict tagging policy at Live Proxies: every cloud resource is tagged with relevant metadata.”

This enables optimal cost allocation and also enhances accountability throughout the organization, Kalvo says. When Live Proxies had a spike in its cloud bills, it used tags to determine the exact source of the spike. A new project was consuming more resources than were estimated.

“This gave us the opportunity to fix the issue on the spot and budget accordingly,” Kalvo says.

Closely examine billing documents

It sounds obvious, but closely examining the bills that come in from cloud providers is an important part of controlling cloud costs that some organizations of IT departments might overlook. In some cases, the use of multiple providers and a variety of cloud services makes this more complex.

“Keeping an eye on the billing information is one of the perfect ways to reduce cloud costs,” says Martin Gasparian, managing attorney at Maison Law. “Many business owners do not bother to check the billing details, even the cloud expenses [that] seem to be higher than expected.”

Companies should ask their vendors to provide a breakdown of the billing for cloud services, Gasparian says. “The analysis of these details will help you to cut redundant services being paid for in the cloud packages,” he says. “Just pay for priority cloud services rather than marketed packages that come with unnecessary but complementary services.”

Deploy storage autoscaling

Autoscaling is a method used in cloud services that dynamically adjusts the number of resources available based on the demand on servers. Organizations can leverage this for services such as data storage to reduce expenditures.

“One way we stay on top of our budget is by autoscaling our services based on current storage needs,” says Michael Sawyer, operations director at travel company Ultimate Kilimanjaro. “This way, storage supply meets our data demand where it is, instead of overspending on storage we do not even need.”

Another way Ultimate Kilimanjaro budgets for cloud storage is by offering allocated storage allowances per department, so those who might need more storage have it at their disposal. “To me, this allocation of budget is no different than other department budgets, especially since the need to monitor funds spent on technology is always important,” Sawyer says.

Conducting audits and shutting down unused resources, especially at off-hours, prevents overspending, Gogia says. “One e-commerce company I worked with was able to cut their bill by a third by implementing autoscaling and addressing inefficiencies,” he says. “Performance improved at the same time as costs declined.”

Given the “extremely dynamic workloads at Plitch, the company uses, autoscaling extensively to spin up and down resources in response to demand, Schaal says. “Every two weeks we also analyze usage statistics to identify areas that could be optimized,” he says. “This level of information is essential because unrestrained cloud spending has a tendency to grow rapidly.”

Create a culture that embraces cost control

To help keep cloud costs under control, it’s a good practice to create a company culture that encourages users, including line-of-business leaders, to be aware of expenses

“Having a cloud-conscious team is often ignored,” says Stefan Chekanov, co-founder and CEO of Brosix, a provider of communications and messaging tools. “The human factor can’t be tracked or measured like plain data, but its impact is just as huge.”

Simply having the right tools isn’t enough, Chekanov says. “In my experience, you have to actively promote a culture where everyone understands the impact of their cloud usage. As a leader, you have to have those transparent cost discussions, especially when it comes to development processes.”

Cloud engineers who are aware of the financial implications of their choices are more likely to be resource-efficient, Chekanov says.