At a glance

A regional logistics operator was spending more each month on a cloud environment that still struggled under seasonal demand. 12th Wonder took operational ownership of the estate, optimized its AWS and Azure environments, automated capacity management and introduced ongoing cost governance. Within two quarters, monthly cloud spend fell by 32%, while peak-season uptime improved from 96.2% to 99.95%.

About the client

The client is a regional logistics and freight forwarding operator running road and warehouse services across several markets. It handles shipments for retail, industrial and consumer goods customers, with volumes that swing sharply around seasonal trading peaks.

Its technology estate supports shipment tracking, warehouse management and customer reporting, all of which are used daily by operations teams, drivers and client-side users. The business has grown steadily through both new contract wins and acquisition, which left it running a mix of systems that had never been consolidated onto a single operating model.

The challenge

The client ran core shipment tracking and warehouse systems on an ageing on premise server estate that had been sized for average demand. Seasonal freight peaks pushed the environment past its limits every year, and outages had started to land in the weeks that carried the most revenue.

An earlier migration into AWS was intended to fix this. It was provisioned generously to reduce migration risk, which meant instances sat well above the workload they actually served. Monthly cloud spend climbed steadily. Peak week performance stayed roughly where it had been before the move.

By the time 12th Wonder was brought in, the finance team could not attribute spend to any specific application, and the engineering team had no reliable way to scale capacity ahead of a known busy period. Every peak season was handled manually and reactively.

What the assessment revealed

We began with a four-week discovery across both the AWS estate and a smaller Azure footprint used for reporting and analytics.

Three findings shaped the programme:

Around 40% of compute capacity was running at consistently low utilization, including several development environments left switched on around the clock

Storage tiering had never been reviewed after migration, so archival data was sitting on high-performance disks

Scaling was entirely manual, which meant capacity decisions depended on one senior engineer being available at the right moment

Our approach

Right sizing and workload placement.

We resized instances and moved predictable workloads to reserved capacity, while using spot capacity selectively for fault-tolerant workloads. Storage was re-tiered based on access frequency, moving infrequently accessed and archival data to more cost-efficient storage classes. Dormant environments were scheduled to shut down outside working hours.

FinOps led cost governance.

Tagging standards were applied across both cloud platforms so that every resource could be traced to an application owner and a cost center. We set up monthly cost reviews with finance and engineering in the same room, supported by dashboards that showed spend by service and by business unit.

Automated scaling ahead of demand.

Three years of shipment data were used to model demand patterns and configure scaling policies around booking volumes and infrastructure signals. Capacity could therefore begin increasing ahead of predictable freight peaks rather than waiting for server utilization to cross a threshold.

24x7 monitoring and response.

Our managed services team took on continuous monitoring with defined escalation paths and documented runbooks for the incident types the client saw most often. Alert thresholds were tuned over the first six weeks to cut noise and surface issues that genuinely needed attention.

The results

32% reduction in monthly cloud spend within two quarters, achieved while transaction volumes grew

Platform uptime during peak season rose from 96.2% to 99.95%, removing the outages that had previously disrupted the busiest trading weeks

Migration investment paid back in seven months against the original business case

Cost attribution now covers every workload, giving finance a monthly view it can plan against

Capacity planning moved from a manual task owned by one engineer to an automated process the whole team can operate

Why it worked

The cost and performance issues came from the same underlying problem: the environment had been moved to the cloud without fundamentally changing the capacity assumptions and operating practices behind it.

Addressing infrastructure sizing, scaling, cost visibility and operational governance as one programme allowed the client to reduce unnecessary spend while maintaining the capacity required for seasonal demand.

The governance model has kept those improvements in place. Twelve months later, the environment remains within the cost envelope established during the programme, while peak-season operations run without the incident bridges that had previously become routine.

Why 12th Wonder

Cloud infrastructure management requires operational, engineering and commercial decisions to work together. With 13 years of experience supporting enterprise infrastructure across the US and GCC, 12th Wonder brings cloud architecture, FinOps and managed operations into a single delivery model.

Our Cloud Infrastructure Management practice supports AWS and Azure environments across assessment, migration, optimization and ongoing managed services. For this client, that meant continuity from initial discovery and redesign through to day-to-day monitoring, governance and support.

That continuity helped ensure the improvements became part of the client's operating model rather than ending with the optimization programme.

12th Wonder is ISO 9001 certified, NMSDC registered and has delivered more than 100 enterprise projects.

Cloud Cost and Performance Turnaround for a Regional Logistics Provider