DIY cloud cost management: The strategic case for building your own tools

Cloud cost management remains a critical CIO priority. With questions around ROI, increasing outlay, and corporate scrutiny on IT cost savings on the rise, CIOs must know not only what contributes to their organization’s overall cloud spend but also how to optimize it. And that’s all before considering the need to fuel new AI initiatives, which can push cloud costs up further.

The dynamic nature of the cloud — and the need for continually optimize operations — often drives requirements unique to a CIO’s enterprise, meaning that even some popular third-party cloud cost optimization tools may no longer fit an enterprise’s specific requirements.

For CIOs who may need to customize their cloud cost information streams or manage a complex cloud estate, do-it-yourself cloud cost management may be the way to go. Here’s a look at why you might want to roll your own cloud cost solution, what makes a successful DIY approach, and how some leading organizations have already done so.

The case for DIY cloud cost management

While commercial cloud cost optimization tools do the job just fine for the average cloud enterprise, several factors are driving enterprises to build their own cloud cost management tooling internally, including:

• Commercial cloud optimization tools can come with high costs, especially as some vendors move to usage-based pricing and charge a percentage of your enterprise’s cost savings.
• Evolving enterprise needs often outpace the product roadmaps of SaaS cost optimization solutions providers.
• Efforts to customize commercial cost management tools can be just as challenging as building a DIY cloud cost optimization solution.

“It gets costly very quick,” says Jim Ducharme, CTO of ClearData, a cloud security posture management (CSPM) and managed detection and response (MDR) SaaS provider for the healthcare industry.

Ducharme cites the complexity of cloud cost optimization tools and “AI washing” in the commercial market as reasons behind his company’s move to DIY. He and his teams tried a few off-the-shelf tools but were never satisfied with the support for linking a cloud resource to a line of business to determine ownership. He also wanted to structure a set of governing policies in which each team must answer questions about the cloud resources they use, the expense associated with their use, and other management options for their resources. This strategy alone — achieved with the DIY approach — was responsible for saving $300,000 when his team dug into expenses surrounding Google Cloud Platform log storage, resulting in his team moving the logs to a lower tier of cloud storage and rightsizing the data they retain to adhere to retention requirements.

Kevin Garcia, vice president of cloud solutions for IT services firm Presidio, sees larger customers with cloud spends in the $10 million per month range for any one cloud service provider (CSP) encountering issues with commercial offerings. For example, the aggregation of billing data, and the act of grouping tags to populate all the attributes that must be applied after data ingestion, can be burdensome on some cloud cost optimization tools, slowing down efforts to react to the spending data. Garcia gives the example of the AWS cURL file, written three times daily.

“You don’t know when it will be written, and you must be ready to consume it. So, [larger customers must] be able to consume that subset of data and run all of the kinds of jobs they want to run on top of it — moving by tags, moving by account ID or business unit, whatever the case might be, in a meaningful way that they can react quickly,” he says.

Here, a DIY approach can help enterprise IT teams be more nimble in making adjustments to their cloud resource use, thereby keeping fine-grained, responsive control over cloud usage before costs have a chance to spike.

Anatomy of a DIY cloud cost management program

For most enterprises, custom cloud cost management strategies will begin by leveraging tools available from CSPs themselves.

Amazon, Microsoft, and Google all provide native tools for monitoring your cloud usage and spend within their respective cloud platforms.

Chris Hennesey, an enterprise strategist with AWS and a former technology CFO for Capital One, advises enterprises CIOs to first consider cloud-native tools from their CSP.

“For AWS, many of the key capabilities are tools available to all our customers, such as AWS Cost Explorer and the AWS Cost and Usage report,” he says. “Recently, AWS also announced Cost Optimization Hub, which centralizes many cloud cost recommendations.”

Other CSP-native tools, such as Microsoft Cost Management for Azure, Power BI, and Looker, can be a foundation for your DIY cloud cost optimization solution. You can then augment these solutions with open source and commercial reporting tools you may already have in-house, such as Tableau.

Beyond these foundational tools, DIY cloud cost optimization requires a strategy that enables your organization to leverage spot instances, reserved pricing, cloud savings plans, and similar strategies to save on cloud storage costs.

Moreover, it requires a culture shift, says Hennessy, stressing the need for a “builder culture” in order for DIY cloud cost optimization to succeed.

Ducharme agrees, adding that teams asking questions about their cloud environment is a necessity. Also, your DIY tool is best served in a culture where iteration is accepted as you add new features to your platform over time, on your schedule.

Hennesey also advises tackling DIY cloud cost management in “somewhat of a constrained way.”

“That could be constraining the number of resources, maybe putting a constraint on time,” he says, offering some “think big, act small” tips for CIOs who’ve put a priority on cloud cost management, including:

• Dedicate enough staff resources and capacity to ensure a viable effort.
• Assign a product owner.
• Follow an agile model to iterate and deliver value over time.
• Set an eight-week development cycle with two-week sprints.
• Have a demo of the proof of concept at the end of the eight weeks.

“I think a CIO can make a minimal investment … to build that out and see the progress they’ve made,” says Hennessey, who also advocates pitting your DIY proof of concept of against a commercial tool. “If you have a desire to build. I don’t think you have to look past that option because we just see plenty of customers pursuing it.”

Presidio’s Garcia believes a key first step to building a DIY cloud cost optimization solution is understanding how your organization manages cloud and cloud costs today versus what’s not being done. He gave the example of a large organization where some departments practiced good cloud hygiene but the same couldn’t be said for three-quarters of the other departments. Building out a governance strategy based on the practices of departments that are getting it right is vital for tracking — and reining in — your cloud budget.

At a minimum, your DIY cloud cost optimization team will require an enterprise architect who understands the technology, says Garcia, who also recommends a financial developer or somebody with financial and data science experience. Granted, that can be a hard combination to find, but they’re out there, he says. You can then start building data lakes and models around your data.

“Having the financial aspect under control is key,” he stresses. The cross-section of financial and developer background becomes important in running queries using BigQuery and other tools and packaging the financial data into stakeholder reports. A data science engineer who understands business stakeholder needs and ties them back to developing the financial features of your DIY cost optimization solution is another recommendation. The team also needs finance engineers — hands-on keyboard individuals — who can employ the governance controls in each CSP.

For smaller companies, Ducharme recommends adding cloud cost management responsibility to the cloud operations team, which he did at ClearData. His cloud ops team already had access to the data and just needed to add governance processes to their duties. According to Ducharme, a small company’s cloud operations team can readily draw on their application performance and resource management skills and experience to support cloud cost optimization better, especially when DIY tools come into play.

Of course, not every IT department may be set up for DIY cloud cost optimization success. For example, when discussing DIY approaches with AWS customers, Hennesey will ask what their balance is between in-house and third-party developers.

“I would say if a company fully outsources a lot of their third-party capacity into a managed service arrangement, I think leveraging a third-party tool would be the best first choice in that regard,” he says.

DIY tools in action

A number of leading organizations have already taken a DIY approach for cloud cost optimization rather than relying on a third-party solution. Here are a few notable examples.

Airbnb: The online property rentals company provides a prime example of optimizing cloud costs using a DIY approach. The company uses the AWS Cost & Usage Report and Amazon S3 Intelligent-Tiering and Savings Plans to cut expenses. Amazon OpenSearch Service is used to manage costs within logging infrastructure. The service provides interactive log analytics, real-time application monitoring, website searches, and more. Airbnb has employed UltraWarm storage for Amazon OpenSearch Service, resulting in a 60% reduction in costs to help fuel sustainable growth.

Formula 1: The automobile racing organization has achieved an 80% reduction in computational fluid dynamics (CFD) simulation time and a 30% decrease in the cost of running workloads by using a combination of Amazon EC2 instances. This has enabled Formula 1 to better accomplish its strategic objectives of increasing competitiveness and unpredictability on the track, as well as delivering a world-class spectacle for fans.

ClearData: The 200-person company optimized cloud costs using the same CSPM it sells to customers. The tool collects cloud inventory from AWS, Azure, and GCP into a central environment and implements a tagging strategy for resources during provisioning. ClearData’s tech stack ensured that appropriate tags remained during deployment. According to Ducharme, his team saved over half a million dollars in cloud costs in the past quarter using its own tool and strategies.