Stopping cloud sprawl: How a marketing company cut costs and improved security

When Kforce began the project, the company’s GCP environment had expanded into a range of services, including Compute Engine virtual machines and Google Kubernetes Engine container orchestration. It also included Cloud Run and App Engine serverless platforms, along with managed services by Cloud Storage and BigQuery.

While these efforts enabled innovation and speed, they also prevented executives from getting a transparent view of the cloud’s architecture health and operational resilience. Lacking the in-house GCP expertise needed to untangle the cloud sprawl, the company’s cloud spending kept rising while financial accountability and governance fell behind.

Workloads in GCP were designed and deployed by multiple teams using different architectural approaches. Core best practices were applied unevenly, such as standardized identity and access management, environment segmentation, resiliency planning and operational monitoring.

Because architectural decisions were distributed across teams, leaders lacked confidence in their ability to assess platform-level risk. Questions about fault tolerance, incident response readiness and operational ownership couldn’t be answered consistently across environments.

Even though the company wasn’t dealing with cloud sprawl in their AWS environment, they didn’t have a FinOps operating model in place. This meant there was no clear way for the engineering, finance and leadership teams to work together on forecasting, governing or optimizing cloud costs. Instead, everyone tackled cost issues as they popped up and relied on ad hoc reviews. Because of this, they often missed out on discounts that could have helped.

Partnering with Kforce, the company ran two assessments: one for GCP and the other for AWS. The GCP assessment focused on architecture, security, best practices and financial analysis. With AWS, the concentration centered on FinOps, with financial analyses run on reserved instances, licensing and resource sizing to find cost savings.

Kforce designed several solutions for the company that balanced near-term risk reduction with long term maturity. After talking with the company about its needs, Kforce crafted solutions that provided leadership with clarity, financial discipline, risk reduction and a sustainable cloud growth operating model.

Kforce proposed formalizing a set of cloud architecture standards to establish clear, repeatable patterns for how environments should be built, secured and operated going forward. This included recommendations to standardize network segmentation between environments, strengthen identity access management controls, and ensure that logging and monitoring were consistently enabled across workloads. A list of security recommendations based on the client’s available tools was created, focusing on urgent steps to lower risk.

Kforce proposed moving from ad hoc cost optimization for AWS to a structured FinOps model that aligned engineering, finance and business stakeholders. This included improving tagging standards and enhancing cost reporting for product and application owners. All recommendations were categorized into immediate, near-term and long-term priorities and explained which tools and controls to use, why they were appropriate and the scenarios where they’d deliver the most value. In addition, regular review cadences were established to evaluate usage, forecasts and optimization opportunities.

Kforce reviewed usage patterns and suggested specific ways to optimize the AWS infrastructure, especially for computing and storage resources that were not being fully used. Recommendations included adjusting instance sizes, reducing extra storage capacity and matching usage habits with cost-effective pricing options like reservations or savings plans.

Across both AWS and GCP, Kforce proposed consolidating findings into a single roadmap for leadership to guide investment and sequencing decisions. Recommendations would be grouped by risk, impact and effort, making it easier for executives to understand tradeoffs and timing. The team also created a cost summary and analysis identifying the primary drivers of costs and inefficiencies. They also made a detailed resource inventory that covered all the infrastructure supporting business-critical applications, including computing, storage and supporting services.

The Kforce team’s discoveries gave the company’s leaders the data-driven insights they needed to reduce their ballooning cloud costs by 14% per month.

• 14% potential reduction in monthly AWS spend was identified through modeled FinOps scenarios, driven by rightsizing, discount instruments and lifecycle policies.
• 60% projected savings from more cost-effective EMR workload processing.
• 25% projected savings from improved object storage and data management practices.
• 43 AWS virtual machines analyzed as part of a single application FinOps assessment, providing a granular view of utilization, cost drivers and optimization levers.

The team replaced uncertainty around risks, inefficiencies and gaps in both the cloud architecture and FinOps with evidence-based insights. Company leaders can now make informed decisions about cloud strategy, instead of relying on assumptions or incomplete data.

The assessments surfaced previously unknown security and governance gaps, ranked by severity for disciplined and defensible risk management decisions. Leadership can now see which issues require immediate attention versus those that could be addressed over time.

The FinOps review turned cloud usage data into clear and practical information, helping both finance and engineering teams see exactly what drove costs and where savings were possible. This common understanding laid the groundwork for more productive conversations about tradeoffs between cost, performance and scalability.

The GCP assessment showed the client where it lacked the expertise and governance needed to operate a platform at scale. This clarity informed broader cloud platform consolidation decisions, reducing operational complexity and improving long-term manageability.