Case Study: Rightsizing for Cost Savings

Cloud costs spiralling out of control? Rightsizing could save you up to 60%.

This case study explores how a UK enterprise slashed cloud expenses by 34% (£115,000 per month) and improved efficiency by aligning resources with actual demand. The process involved:

• Auditing resources: Identifying unused or over-provisioned servers, reducing waste.
• Auto-scaling: Dynamically adjusting resources based on real-time demand.
• Monitoring tools: Custom dashboards tracked savings and highlighted inefficiencies.
• Expert support: Partnering with specialists for tailored solutions and implementation.

The result? Lower costs, better performance, and a more streamlined cloud infrastructure. If your cloud bills are rising, rightsizing could be the solution you need.

Cloud Cost Problems Before Rightsizing

High Cloud Costs and Resource Waste

When the enterprise embarked on its cloud migration journey, the promise of cost savings quickly turned into spiralling expenses, with no visible improvement in performance.

The main culprit? Overprovisioning. Teams were assigning far more CPU, memory, and storage resources than applications realistically needed. Take one instance as an example: it was allocated 4,096 MB of RAM, even though its workload used only a fraction of that. Research suggests this isn't an isolated case – up to 84% of cloud instances may be overprovisioned[4].

On top of this, the company faced the issue of zombie servers – virtual machines that had outlived their purpose but were still running and racking up costs[1]. Monitoring was another weak point. Teams primarily tracked CPU usage, often neglecting other critical metrics like memory usage or network performance[2]. Without a complete picture of resource consumption, identifying inefficiencies or pinpointing spending hotspots became almost impossible.

The scale of waste was eye-opening. Audits revealed hundreds of underutilised services and applications. For example, one organisation found it was using 127 different SaaS applications, with 42% of licences either unused or underutilised. They also discovered 23 duplicate tools performing the same tasks[3]. Another case demonstrated the potential for savings: downsizing 1,465 AWS instances slashed costs by 34%, saving £116,000 every month[5].

This mismanagement didn't just inflate costs – it also paved the way for operational headaches.

How Waste Affected Operations

The inefficiencies in resource allocation went beyond financial strain, creating operational challenges that hindered performance and stifled growth. Addressing these issues became critical to establishing a more effective rightsizing strategy.

One major consequence was the inability to reinvest in innovation. The money spent on idle resources could have funded new projects or essential upgrades[2]. Instead, the enterprise was paying premium prices for cloud services while struggling to allocate resources towards strategic improvements.

Operational complexity also grew exponentially. Fragmented and duplicated resources increased the workload for maintenance teams and slowed down deployment cycles[3]. Development teams found themselves bogged down with infrastructure management tasks, detracting from their ability to deliver business value. The lack of standardised processes across departments exacerbated the problem, as similar applications were often run on separate platforms, leading to integration headaches and extra training requirements.

Performance issues became another pain point. Misaligned resources meant some applications faced bottlenecks, while others were over-resourced and underutilised. This imbalance led to slower response times and more frequent troubleshooting efforts.

Scaling was another area of difficulty. Overprovisioning made it challenging to differentiate between routine operations and genuine demand spikes. This confusion often resulted in either inadequate scaling responses or even more overprovisioning – perpetuating the cycle of waste.

Perhaps the most damaging outcome was the reactive approach to cost management. Instead of implementing strategic, well-planned optimisations, the enterprise often resorted to abrupt cost-cutting measures when bills became unmanageable. These quick fixes disrupted operations and failed to address the root causes of inefficiency[2]. This short-term mindset is a common industry problem, where delaying optimisation efforts only leads to deeper operational and financial challenges.

Without aligning resources to actual needs, organisations face a dual struggle: keeping costs under control while maintaining strong operational performance. It was clear this enterprise needed a solution to break free from this cycle.

Choosing and Setting Up the Rightsizing Platform

Platform Selection Requirements

After grappling with high costs and wasted resources, the enterprise realised that finding the right platform was crucial to their cost-saving efforts. They needed a solution that could meet their complex needs without disrupting existing workflows.

Scalability was a top priority. The platform had to handle large, fluctuating workloads across departments with varying usage patterns. With a diverse infrastructure consisting of numerous instances, the solution needed to operate effectively at this scale.

Integration was another key consideration. The platform had to offer strong API support and work seamlessly with their existing DevOps tools, such as CI/CD pipelines, Terraform, Ansible, and monitoring systems. These tools were already deeply embedded in their workflows, and teams needed to continue using them without interruption.

Unified visibility across AWS, Azure, and on-premises environments was also essential. The enterprise wanted to avoid juggling multiple tools for different providers. The platform had to replace manual processes with automated, AI-driven optimisations to ensure efficiency.

Security and compliance were non-negotiable, especially given the requirements of UK data protection regulations. The platform needed to support detailed access controls, maintain audit trails, and provide financial reporting in GBP to ensure accurate budget tracking and reporting to the board.

As part of the technical evaluation, the enterprise tested each platform's ability to deliver detailed cost visibility. Teams needed to pinpoint exactly where spending was occurring - down to individual resources and business units. This level of granularity was crucial for making quick, cost-effective adjustments.

Custom Setup and Implementation Support

To navigate the complexities of platform selection and implementation, the enterprise partnered with Hokstad Consulting. Known for their expertise in DevOps transformation and cloud cost management, Hokstad Consulting played a pivotal role in aligning the platform with the enterprise's operational needs.

Hokstad began with a comprehensive cost audit, mapping every cloud resource to its corresponding business function. This process highlighted integration opportunities and areas where automation could deliver immediate results.

The platform was then configured to match the enterprise's operational patterns. Hokstad introduced a custom tagging system that aligned with the company’s internal cost-allocation methods. This tagging extended beyond financial tracking to include operational details like environment type, criticality level, and maintenance schedules. This approach ensured departments could monitor their cloud spending with precision.

Automated policies were designed to optimise resource usage. These included resizing instances based on utilisation, scheduling non-critical workloads during off-peak hours, and flagging unusual spending patterns. The policies were tailored to business-specific needs, such as maintaining peak performance for customer-facing systems during UK business hours while enabling aggressive cost-saving measures outside these periods.

Custom dashboards were created to cater to different stakeholders. Technical teams received detailed metrics, while finance teams got cost breakdowns in GBP. High-level summaries highlighting savings and optimisation opportunities were provided to executives.

To tackle integration challenges, Hokstad developed custom scripts to bridge gaps between legacy systems and the new platform. A phased rollout ensured stakeholders were gradually introduced to the system, helping to build confidence.

Training was another critical component. Hokstad delivered tailored sessions designed specifically for the enterprise's setup and use cases. These sessions not only covered how to use the platform but also explained the reasoning behind key configuration decisions, empowering teams to make the most of the new tools.

A structured review process was established to keep the platform aligned with evolving business needs. Monthly assessments were scheduled to evaluate its performance, update policies, and incorporate feedback from technical teams. This ongoing review ensured the platform remained dynamic rather than becoming outdated.

Hokstad’s No Savings, No Fee model further reinforced their commitment to the enterprise’s cost reduction goals. This approach ensured the platform configuration was always focused on delivering measurable financial outcomes. With the platform fully operational and optimised, these efforts began translating into noticeable savings.

Rightsizing - lets make it easy | The Keys to AWS Optimization | S7 E7

Implementation Steps and Actions Taken

Once the platform was set up and operational, the enterprise transitioned to the practical implementation phase. This phase prioritised auditing resources, introducing auto-scaling, and establishing monitoring systems to reduce inefficiencies and control costs.

Resource Audit and Waste Removal

The first step was a thorough resource audit using automated discovery tools to catalogue all cloud assets across AWS, Azure, and on-premises environments. This audit revealed a range of inefficiencies that had built up over months of unchecked provisioning.

The findings included zombie servers that consumed resources without purpose, duplicate storage volumes holding identical data in multiple locations, and outdated database instances that were still active long after their associated projects had ended.

To address these issues, the team analysed six months of historical usage data, focusing on CPU, memory, storage, and network utilisation. Resources with minimal or no recent activity, as well as those excessively overprovisioned, were flagged for downsizing or removal.

The clean-up process involved shutting down unused and orphaned resources, consolidating duplicate services, and resizing overprovisioned instances to better fit actual demand. For instance, migrating workloads from oversized, underutilised instances to smaller, cost-efficient ones resulted in a 55% reduction in annual costs for certain workloads[4].

The audit also identified redundant software licences and duplicate applications, which were promptly addressed to reduce expenses further.

Before making any changes, the teams carefully reviewed each resource's business context to ensure that redundancy, compliance, and peak load requirements were not compromised. They also implemented acceptance testing and service level agreement (SLA) monitoring to maintain system performance and reliability throughout the optimisation process.

Auto-Scaling and Resource Scheduling

The next stage introduced dynamic resource allocation, allowing the infrastructure to scale resources up or down based on actual demand. Auto-scaling groups and policies were configured to automatically adjust compute and storage resources using real-time workload data.

These scaling policies were tailored to typical UK business hours. Resources were scaled up during peak times - generally from 09:00 to 18:00 GMT - and scaled down during quieter periods. This approach ensured sufficient capacity during high-demand times without overprovisioning.

For non-production environments, automated scheduling proved especially effective. Development and testing systems were programmed to shut down outside standard business hours, providing immediate cost savings without affecting productivity. Shutting these systems down over weekends added further savings, as most development work occurred during weekdays.

The scaling policies included built-in buffers to handle unexpected demand spikes, ensuring smooth adjustments without performance disruptions. Gradual scaling mechanisms prevented sudden, disruptive changes in resource allocation.

Additionally, the auto-scaling system was integrated with existing DevOps tools like CI/CD pipelines. This ensured seamless resource provisioning for development workflows. Infrastructure management tools such as Terraform and Ansible were updated to reflect the new scaling parameters, maintaining consistency across operations.

Monitoring Dashboards and KPIs

The final step focused on creating robust monitoring systems to track resource usage and manage cloud spending effectively. Custom dashboards provided real-time visibility into key performance indicators (KPIs), helping teams evaluate the success of optimisation efforts.

Metrics such as CPU and memory usage were used to measure resource efficiency, while storage consumption data highlighted areas for further improvement. Cost-per-workload calculations showcased the financial impact of the changes. These insights were displayed on dashboards designed for different audiences:

• Technical teams received detailed metrics on resource performance and utilisation patterns.
• Finance teams accessed cost breakdowns in GBP, aligning with internal budgeting processes.
• Executives viewed high-level summaries of savings and optimisation opportunities to guide strategic decisions.

One critical KPI, the Effective Savings Rate, tracked realised savings instead of merely projected benefits. This ensured that cost reductions were reflected in actual invoices, providing concrete proof of the programme's success[2]. Monthly reviews showed consistent savings growth as optimisations were implemented.

Real-time alerts flagged unusual spending or resource usage patterns, enabling teams to address issues before they impacted budgets. Automated notifications were sent when resources exceeded utilisation thresholds, prompting reviews for further optimisation.

The monitoring systems also supported continuous improvement. Regular dashboard reviews helped identify new inefficiencies as business needs evolved, ensuring resources stayed aligned with current demands. This ongoing visibility was key to maintaining cost control and preventing waste from creeping back in.

Results and Cost Savings

Rightsizing brought about immediate financial benefits and operational improvements.

Cost Savings and Performance Improvements

The rightsizing initiative led to a 34% reduction in monthly costs, saving approximately £115,000, while average CPU utilisation jumped from under 15% to over 50% across key workloads. This shift significantly reduced idle capacity, cutting down on unnecessary expenses and improving application responsiveness during peak demand.

Performance metrics reflected these changes, with deployment times decreasing by 20%. This was achieved through streamlined infrastructure and fewer resource-related issues, allowing teams to speed up development cycles and bring new features to market faster. Resources were better aligned with actual needs, avoiding the inefficiencies of over-provisioning.

One notable success involved a legacy analytics server. Previously running on an oversized m4.large instance costing £1,200 annually, it was migrated to a t3.medium instance at £540 per annum, resulting in a 55% cost saving without any loss of performance [4].

This experience mirrors that of a UK medical provider, where rightsizing efforts delivered a return on investment within the first month [6]. Similarly, this enterprise saw immediate results that justified the upfront investment in platform setup and implementation.

Beyond direct cost reductions, the optimisation freed up funds for innovation. Development teams were able to access additional resources for experimentation and new feature development, while finance teams benefited from more predictable and accurate cloud spending aligned with business needs.

These advancements not only reshaped spending habits but also laid the groundwork for ongoing improvement and efficiency.

Ongoing Management and Security

Building on these initial savings, the enterprise adopted robust monitoring practices to maintain cost efficiency and strengthen operational security. Real-time dashboards tracked utilisation, costs, and performance metrics, offering immediate insights into resource usage patterns.

Monthly reviews were introduced to evaluate new workloads and adjust resources as business needs evolved. Automated alerts flagged unusual spending or resource use early, enabling teams to address potential issues before they escalated into larger problems.

Security also saw enhancements alongside cost optimisation. Decommissioning unused resources reduced the attack surface, cutting down the number of endpoints requiring oversight [3]. Every change underwent a thorough security review to ensure compliance with internal and external requirements.

The monitoring system integrated automated security scans into daily operations, ensuring cost savings did not compromise security or regulatory standards. Additionally, tracking realised savings - focusing on actual invoice reductions rather than projections - became a key part of the process [2], providing clear evidence of success and uncovering further opportunities for improvement.

To sustain these gains, the enterprise established knowledge-sharing initiatives. Regular training sessions and updated documentation ensured teams stayed informed about evolving cloud cost management techniques and pricing models, helping to maintain long-term control.

For organisations looking to embark on similar rightsizing projects, experts like Hokstad Consulting offer tailored support in DevOps transformation, advanced cloud cost management, and rightsizing automation. Their expertise ensures sustained cost savings, optimised deployment cycles, and compliance across various cloud environments.

Key Lessons and Best Practices

The process of rightsizing has provided several insights that can help other organisations aiming to optimise costs. These lessons are drawn from both the challenges faced and the successes achieved during implementation.

Implementation Lessons

• Keep an eye on all metrics – memory, network throughput, storage, and I/O. A resource might seem underutilised when viewed through CPU usage alone, but it could be running at 85% memory capacity or facing network bottlenecks that justify its current size.

• Focus on actual invoice savings rather than dashboard projections. Predicted savings don’t always translate into reality due to pricing complexities, reserved instance commitments, or billing inconsistencies. Tracking real savings adds an extra layer of assurance for ongoing optimisation efforts [2].

• Understand the context of each resource before cutting capacity. Unused capacity might be there for seasonal demand, compliance, or failover purposes. Investigating the original purpose of resource allocation helps avoid costly errors and potential disruptions [2].

• Automate routine downsizing but rely on engineers for complex optimisations. Automated tools can yield savings of 5–15%, but engineering-led efforts like query tuning, caching, and architectural adjustments can deliver 20–30% or more. These require expertise and a tailored approach [2].

• Adopt automated discovery tools to find forgotten resources. Identifying unused subscriptions and zombie servers revealed budget-draining resources that provided no value [3].

Maintaining Cost Control

Once implementation is complete, it’s crucial to maintain the gains with structured cost management strategies.

• Perform quarterly audits to stay aligned with changing workloads. A consistent process - reviewing usage, identifying opportunities, making adjustments, measuring savings, and refining strategies - ensures optimisation efforts remain effective [2].

• Combine automated monitoring with manual reviews. Automation helps with repetitive tasks, but human oversight ensures recommendations fit the operational context, enabling better decision-making [2].

• Track acceptance and rollback rates to highlight gaps. Monitoring these metrics can reveal areas where additional training or process adjustments are needed [2].

• Leverage rightsizing data during vendor negotiations. For instance, a Fortune 500 company found it had 500 Adobe Creative Suite licences but only 89 active users in the past 90 days. This insight helped renegotiate terms, saving £250,000 annually [3].

• Encourage open communication across teams. Highlighting how rightsizing aligns capacity with demand while improving performance and reducing costs can help address concerns about potential service issues [2].

Conclusion: Results from Cloud Rightsizing

This case study showcases how a thoughtful approach to rightsizing can shift cloud operations from being a financial drain to a competitive edge. The enterprise initially grappled with skyrocketing cloud expenses that showed no improvement in performance. By employing a detailed strategy that blended automated discovery tools with expert engineering, the organisation achieved cost savings of 30% to 60%, delivering immediate returns without compromising performance.

But this initiative wasn’t just about cutting costs. It also enhanced operational efficiency by improving scalability and simplifying processes, creating an environment that supports both innovation and growth [3][7]. These operational improvements laid the groundwork for implementing expert cloud cost engineering practices.

Hokstad Consulting's expertise in cloud cost engineering played a pivotal role in these achievements for UK enterprises. Their tailored strategies addressed challenges like complex cloud infrastructures and developer productivity, making a significant impact.

One of the key takeaways from this success story is the importance of focusing on realised savings rather than projections, understanding the reasons behind resource allocations, and establishing ongoing practices to maintain efficiency [2]. By adopting this comprehensive approach, the enterprise has set itself up for long-term cost control and operational success.

For UK companies struggling with similar cloud cost issues, this case study underscores the tangible financial and operational benefits of rightsizing. The real secret lies in combining the right tools with expert guidance to strike an effective balance between cost, performance, and security - ultimately boosting overall efficiency and creating a solid foundation for future growth.

FAQs