At a glance
A leading global martech platform needed to lower Amazon Web Services (AWS) cloud costs after years of growth through mergers and acquisitions. SoftServe ran an architecture assessment and billing analysis to find where cloud spend could be reduced. The engagement replaced static, anticipated-need resource provisioning with auto-scaling that adjusts to real-time demand. Within 9 months, the client began saving $200,000 to $300,000 per month in AWS costs. The initiative delivered a 297% ROI. The client and SoftServe continue to expand the partnership into data platform, DevOps, and Kubernetes optimization work.
Client background
The client is one of the world's largest martech companies, built through a series of mergers and acquisitions into a market leader. It is privately held and operates an integrated platform that lets marketers manage every part of a marketing program in one place, including the rollout of complex, personalized customer experiences. Its customers include leading brands across multiple continents.
Business challenges
The client's digital marketing platform runs entirely on AWS, and years of mergers and acquisitions had left its cloud footprint provisioned for anticipated demand rather than actual usage. That mismatch drove unnecessary spend and made it harder to scale cost-effectively as the business grew. The client needed a shift from static provisioning to a model that adapts automatically to real demand, while keeping existing systems fully maintained and operational. Specifically:
- Cloud resources were provisioned based on anticipated need, not real-time demand
- Post-merger infrastructure sprawl made cost visibility and control difficult
- Existing systems required continuous maintenance without disrupting operations
- The client lacked a clear, prioritized view of which architecture components drove the most cost
Activities & Solutions
SoftServe delivered a value-based AWS cost-savings proposal, and moved the client's architecture from manual, anticipated-demand provisioning to auto-scaling that eventually saved $200,000 to $300,000 a month. The proposal covered automation testing, a CI/CD proof of concept, an Oracle-to-Snowflake data migration, a message queue (MQ)-to-Amazon Simple Queue Service (SQS) migration, Terraform-based infrastructure-as-code optimization, and application re-architecture.
A joint SoftServe-client team of DevOps engineers, senior software engineers, big data architects, and solution architects carried out the initiative in a defined sequence:
- Analyzed billing data from AWS Cost Explorer, monitoring data from AWS CloudWatch, and usage trends in Grafana
- Ran an architecture assessment using a quality attribute workshop and the architecture tradeoff analysis method
- Applied AWS Well-Architected Framework principles, AWS's official best-practice guidance for cloud architecture, to identify cost-saving redesigns
- Analyzed underused resources on Kubernetes, a container orchestration platform, with an automated toolset to eliminate waste
- Tied application business goals to a SoftServe reward model based on optimized cost per unit of client output
- Delivered a discovery report with prioritized cost-saving initiatives and savings estimates accurate to within plus or minus 5%
This cloud optimization initiative also opened the door to related workstreams, including maintenance and support, UI/UX modernization, CI/CD modernization, automated-testing modernization to support a migration to Google Looker, and data platform and data-lake modernization.
Value delivered
SoftServe's cloud optimization program cut the client's AWS costs by $200,000 to $300,000 every month since January 2022 and delivered a 297% ROI. Auto-scaling replaced static provisioning across core workloads, giving the client's infrastructure the ability to expand and contract with real demand instead of forecasted need. The client now has a clear, cost-mapped view of its architecture and applies auto-scaling to data analytics clusters as an ongoing operating practice, not a one-time fix.
| Metric | Before | After |
| Resource provisioning model | Static, based on anticipated demand | Dynamic auto-scaling based on real-time demand |
| Monthly AWS cost savings |
None tracked against optimized baseline | $200,000–$300,000 saved per month since January 2022 |
| Return on investment |
Not measured | 297% ROI |
| Cost visibility |
Limited post-merger visibility | Clear, prioritized view of highest-cost architecture components |
SoftServe and the client are now extending the model to a data warehouse (DWH) re-platforming project to lower license costs, further Kubernetes resource optimization, and cost-saving measures applied to non-production environments during off-hours.
Turn your AWS spend into a growth advantage. As an AWS Premier Tier Partner, SoftServe engineers the architecture that lowers cost today and scales with tomorrow's demand. Let's talk!
