Many businesses have learned the hard way that migrating to the public cloud doesn’t always result in the cost savings they hoped for. In fact, it’s predicted that through 2020, 80% of organizations will overshoot their cloud IaaS budgets.1
From our experience with thousands of cloud customers, we’ve seen that a majority of wasted cloud spend stems from a handful of causes. In this article, we’ll share six of the most common causes, along with best practices to optimize cloud costs for the future.
Six common causes of wasted cloud spend
1. Overprovisioned resources
One of the most common causes of wasted cloud spend is overprovisioned resources. Oftentimes developers will request more resources than they really need, to ensure they’ll have enough to manage their workload. Understandable, but over time, this practice leads to a significant amount of wasted cloud spend on resources that weren’t needed.
Here are a few tips to help prevent overprovisioning:
- Attain and provide visibility: It’s almost impossible to accurately measure and manage cloud costs without complete visibility of your cloud infrastructure. Support and visibility can be a two-way street if you give your team access to a cloud management platform. You can learn more about how you can use CloudHealth to view costs by dynamic business groups (department, budget owner, project, etc.) here.
- Establish showback/chargeback: When teams have direct financial responsibility for their actions, behavior changes rapidly. Many companies start with a showback approach (which involves sending an invoice to teams and departments showing their spend, but not actually charging them) before eventually implementing true chargeback.
- Implement policies: In order to keep costs under control, the best practice is to implement financial management policies that will help identify which lines of business, cost centers, or projects are accountable for driving up costs, and will alert you when costs unexpectedly spike. There are several different policies you can set, including a policy that shuts down systems during non-production hours, a policy that sends alerts when overall spend is nearing budget limits, or even more specifically, an alert if total Amazon S3 costs increase by more than 10% in a day.
Even if you’ve overprovisioned resources, you can prevent additional costs by rightsizing your cloud environment. Rightsizing is the process of analyzing the utilization and performance metrics of your infrastructure, determining whether or not they’re running efficiently for what you’re paying, and what actions you should take to improve by modifying the infrastructure as needed (upgrading, downgrading, terminating). This can be done for compute, database, storage, containers, and many other infrastructure types.
Rather than manually checking every instance individually to ensure it’s rightsized for optimal cost and performance, a cloud management platform will be able to provide you with the metrics and recommendations you need to make accurate rightsizing decisions.
2. Unused assets
Unused assets are virtual machines (VMs) or instances that are being paid for, but not in use for the time it’s being paid. This can be a side-effect of overprovisioning, but can also be due to the following:
- Unattached virtual server disk infrastructure
- Aged snapshots
- Disassociated IP addresses
- Zombie assets
- Running systems during non-production hours
For detailed best practices to prevent paying for unused or idle cloud resources, check out our in-depth guide.
3. Data in the wrong place
Amazon’s S3 storage service is one of the most profitable divisions within Amazon—and there’s a good reason for this. Most AWS users take advantage of Amazon’s Simple Storage Service because of its instant access, high scalability, and guaranteed uptime. However, not enough users take advantage of the Infrequent Access and Glacier storage classes.
Amazon actually offers two Infrequent Access storage classes—Standard Infrequent Access, where data is stored in a minimum of three availability zones, and One Zone Infrequent Access, which stores data in a single availability zone and is about 20% less expensive than Standard Infrequent Access. If you’re storing data that isn’t necessary for disaster recovery, you can reduce multicloud costs by transferring non-archive data to the One Zone class.
This is just one common example, but the bottom line is, make sure your data is in the most optimal place. If you’re looking for more information, we outlined AWS, Azure, and GCP cloud services and storage options in this article and created an in-depth guide to Amazon cloud storage in this guide.
4. Failing to take advantage of cloud pricing discounts
Leading public cloud providers offer a range of commitment and discount options so customers can customize how they pay for their cloud infrastructure. However, many customers don’t take advantage of the cloud pricing discounts that providers have to offer.
AWS states on their pricing page that you can save up to 75% when you use Reserved Instances, and Azure says you can save up to 90% compared to pay-as-you-go prices when you use Azure Spot Virtual Machines. Even if you don’t fit the requirements to reach the maximum level of savings, there are still significant cost reductions you can achieve when you take advantage of your cloud provider’s discount options.
But we get it—especially if you’re using more than one cloud provider, it can be difficult to know when and where to leverage a cloud pricing discount. To help simplify it, we created a comparison of AWS, Azure, and GCP cloud reservations and pricing discounts so you can more easily see where discounts are available for each provider.
5. Lack of automation
No matter what level of presence your business has in the cloud, governing the environment is complex, especially considering the growing number of users that can deploy workloads in the cloud, and the speed at which they can do it.
To stay on top of cloud costs, leading organizations rely on automation. Automated cloud policies can increase the number of violations detected, accelerate responses to policy violations, and ultimately, reduce wasted cloud spend.
Examples of automated cloud policies include:
- A notification when costs are projected to exceed a monthly budget
- Suspending the launch of a Virtual Machine if its CPU capacity exceeds a certain level
- Terminating a Virtual Machine with unauthorized open ports
- Revoking access to an account that’s logged in from a non-conforming IP address
- Scheduling on/off times for non-production assets
- Switching Reserved Instances between Regional and Availability Zone scopes
- An alert about data within frequent access storage volumes that haven’t been accessed recently
Any organization that has a cloud management platform with automation capabilities can implement automation to reduce wasted cloud spend. To learn more about how CloudHealth can help you get started, see our cloud governance and automation solutions page.
6. Siloed decision-making
The last common cause is not a technical one. The biggest challenges for organizations embracing the cloud often involve people and processes, not technology.
If you’ve read any of our resources before, you’ve probably heard of a Cloud Center of Excellence (CCoE). Now more than ever, every business with a presence in the cloud should have a CCoE—a cross-functional team tasked with supporting and governing the execution of the organization's cloud strategy.
An effective CCoE brings together key stakeholders across the business, including security, finance, governance, operations, IT and engineering. Thyle Carroll, the Director of Delivery of Enterprise Services and Cybersecurity at The Financial Times emphasizes that “the success of an organization’s digital transformation in the cloud relies on all parts of the business.” In sum, don’t make decisions about the cloud in silos.
Work with trusted partners
You might feel overwhelmed by the end of this list. Understandably, it can be a lot to take in, especially if you’re just starting out on your cloud journey. But the good news is that there are tools, services, and systems that can help you get to where you want to be.
Take advantage of leading cloud cost management platforms like CloudHealth, and our global partner community that has helped thousands of organizations achieve their goals in the cloud. You can learn more about how to best leverage cloud management solutions to take control of your cloud strategy and optimize your cloud costs in an article from Mission, a CloudHealth partner and AWS Premier Consulting Partner.
It’s important to remember that the best practices we’ve listed are not meant to be one-time activities, but ongoing processes. Because of the dynamic and ever-changing nature of the cloud, cost optimization activities should ideally take place continuously.
“We recommend not approaching cloud optimization like a one-time project, but rather, a regular, ongoing process of refinement and improvement.” — Jonathan LaCour, CTO, Mission
For more information about optimizing your cloud costs and building a successful cloud financial management practice, see our in-depth guide: Building a Successful Cloud Financial Management Practice
1. Gartner, How to Identify Solutions for Managing Costs in Public Cloud IaaS, Brandon Medford and Craig Lowery, 22 January 2018