Mastering the Cloud Post-Migration: Best Practices for Managing Costs and Driving ROI 

By Siva Padisetty, Chief Technology Officer, New Relic 

The surge in AI adoption has brought a corresponding uptick in cloud migration to tamp down on AI’s spiraling costs. The expenses for enterprises to acquire, clean, and manage their data to leverage AI turned out to be high, but have skyrocketed even more recently. In addition, the significant processing power and energy needed to power AI have piled on additional cloud computing and hardware costs. Desperate to push back on rising costs, organizations are embracing a cloud-first approach, ditching incumbent on-prem or private configurations for public (or at least hybrid) cloud configurations to control the vast amounts of data required for AI.  

The necessary hardware, technical resources, and financial overhead to build and maintain AI/ML infrastructure on-premises are too much for most organizations. By contrast, public cloud providers have made the cloud more appetizing by offering pre-trained models, specialized hardware (like GPUs or TPUs), and managed services to simplify development and deployment. According to Gartner, 90% of organizations will adopt a hybrid cloud approach by 2027, with spending on public cloud services expected to reach $723.4 billion this year.   

However, there are more steps to a successful migration than may appear at first glance. While the migration is transitory, switching to a cloud-centric model is a permanent, ongoing commitment. Organizations must ensure a comprehensive post-migration strategy is in place before they start. Fidelity to cloud migration post-transition best practices is emerging as the primary determining factor of whether the migration effort will succeed. Many companies overemphasize the cloud migration itself, at the expense of the permanent changes required to uphold the strategy and manage the challenges that come with it, like talent needs, changing policies, and cost optimization. 

The Expertise 

The biggest misstep by organizations is to assume that migration is “complete” once the transition phase is done. Instead, critical post-transition steps are essential to maintaining success and managing spending. For starters, talent must be fine-tuned to support the new model on an ongoing, permanent basis. Organizations must determine whether they should upskill current employees or bring in new talent to complement existing teams and manage the cloud effectively. Prior visibility from existing team members is always helpful. Still, a fresh perspective from newly ingrained colleagues with varying skill sets can help the team identify new opportunities they may not have otherwise explored. Nevertheless, the input and expertise of specialized personnel, including cloud architects, cloud engineers and developers, cloud security specialists, DevOps engineers, cloud operations managers, data engineers and data scientists, cloud compliance and government experts, cloud product managers, and cloud consultants, is required.  

Specialized expertise will also be needed to confront the significant financial risks encountered post-migration. Establishing a designated FinOps team to implement cloud financial management practices for real-time cost estimates and detection of anomalies will create vigilance to address the most significant financial risk: overspending. This team should comprise members from finance and engineering who can uniquely evaluate everything, including the cloud services, how billing is handled in the consumption model, and how engineering uses the services. The team’s work is a constant effort to find opportunities for savings, identify and implement optimizations, and normalize cost data to suit the needs of the business. 

Governance and Compliance 

The best combination of talent and teams is essential to achieving success, but is insufficient without the right governance and control processes. From a process and financial standpoint, these help monitor performance, analyze risk, and determine resource management. In terms of management, empowering teams to be more agile with proper control improves productivity and innovation. However, the bottom line is that companies risk non-compliance, overspending, security breaches, and mismanagement of cloud environments if they fail to adopt these.  

Optimization 

Once the building blocks of a strategically crafted team and processes are in place, continuous cloud optimization is critical, especially regarding spending and ROI. Even with the firm foundation of a fine-tuned team and strict processes, lack of visibility into cloud systems can lead to overspending and hinder the ability to forecast potential costs.   

Today, organizations should employ full-stack monitoring, or the ability to see everything in the tech stack that could affect the customer experience, as the gold standard for providing enterprises with a holistic view and deeper insights into their cloud models. And while AI is behind the rise of cloud costs, it has also led to the development of new AI-enabled capabilities that expand the insights full-stack monitoring can provide. Post-migration, these are crucial insights around where to optimize cloud spend, including resource utilization monitoring, cost allocation, and cost anomaly detection.  

By monitoring and tracking, companies can better understand resource utilization to easily identify underutilized resources and mitigate inefficiencies that lead to overspending. They can also run cost estimates and gain a detailed view of cost allocation to help identify areas contributing to rising costs and better align them with business objectives. Spotting issues and unusual spend patterns helps alert teams to quickly resolve problems indicative of underlying problems that could ultimately lead to more costly complications. As a result, companies that leverage these tools are better positioned to perform ongoing, continuous optimization post-migration with much greater accuracy and granularity. 

While the transition to the cloud is a significant milestone for companies in the post-AI era, the real challenge lies in maintaining success post-migration. Continuous cloud optimization, backed by strategic upskilling and establishing a FinOps team, and in tandem with strong governance processes and monitoring, is critical to effectively managing resources and minimizing overspending. These strategies allow companies to better align their cloud strategies with business objectives, ultimately optimizing cloud costs and driving ROI.