Case Study Sample

Title: How Vertex Technologies Cut Cloud Costs By 43%

Executive Summary 

Company:Vertex Technologies 

Industry:B2B/SaaS for project management software

Challenge:Cost of maintaining cloud infrastructure growing much more faster than annual revenue

Solution:CloudOptimise's AI-powered cost optimisation platform

Results:

•43% reduction of annual cloud expenditure ($847,000 annual savings).

•Zero degradation performance.

•12-hour ROI analysis.

•Payback period:2-3 months.

The Company Overview 

Vertex Technologies provides project management software for up to 12,000 companies worldwide.This platform helps departments of various companies coordinate complex projects into accessible data.

Founded in 2018,Vertex grew so rapidly that it's user count doubled year after year due to shipping products being prioritised over optimised infrastructure by staff engineers.Development of new features were also accelerated due to rising customer demands.

By late 2024,Vertex operated 2,400 instances across AWS. Monthly infrastructure bills rose to $200,000 and worse still, costs skyrocketed up to 15% while revenue grew by 8% only. The math didn't just comprehend.

“We were spending 2× much more than our annual revenue,” says Vertex CTO David Park.”Every new customer should have improved our margins but instead,cost of cloud infrastructure ate our profits.”

The problem extended much more than expected.Vertex's engineering team couldn't even point out the services that had consumed almost all the finances.When the executives asked where the money was put in,the infrastructure team shrugged because of how AWS bills listed thousands of line items with little or no actionable insight.

Something had to be done immediately.

The Problem:Death By A Thousand Cloud Instances 

Vertex's cloud architecture had metastasised organically over the last six years.Each engineering team made use with whatever resources they had at their disposal due to no central oversight,lack of cost accountability and no one thinking holistically about efficiency in infrastructure.

The result turned out to be predictable waste:

Why?

Overprovisioned instances:Background development which were supported by production-grade servers ran 24/7 even when no one logged in.

Zombie resources:Some forgotten instances and one-off projects were still running behind and incurring alot of costs.

Inefficient storage:Vortex paid for IOPS they're never used because of how databases were configured for maximum performance instead of the actual query loads.

Redundant backups:Multiple backup systems overlapped due to data being backed up several times across different servers.

Poor instance sizing:Developers wasted 40-60% of provisional capacity by choosing larger instance types “to be safe”.instead.

Park's team knew waste existed but couldn't identify it due to third party quotes of $150,000 being kept aside for monthly audits.

“AWS Cost Explorer showed aggregate expenses instead of optimisation opportunities that would have given us some answers,”Park explains.”Every week we delayed meant another $46,000 gone through the drain.”

How CloudOptimise AI Was Discovered 

Park learnt about CloudOptimise AI through a colleague he bumped into at a colleague. Later on,CloudOptimise AI pitch deck was sent to Vertex accounts team. The pitch deck clearly stated CloudOptimise as an AI tool used to analyse cloud infrastructure,identify wastage and recommend specific optimisations within a few hours.

Already desperate for relief,Vertex account team approved a trial because this could be their only chance of survival.

The whole onboarding process surprised Park because of how CloudOptimise AI made use of only read-only AWS access instead of code changes and agent installation.

By morning,the AI had already scanned Vertex infrastructure and displayed key findings on CloudOptimise's dashboard that would have taken consultants weeks to compile.

Key Analysis That Led To The Discovery Of $847,000 Hidden In Plain Sight

CloudOptimise AI identified seven major problems;

1:Overprovisioned compute (312,000): 143 instances were sized inappropriately due to the average utilisation.11% recommendation and the right size were needed to match the actual usage patterns.

2:Total wastage of resources ($187,000): These systems were meant to be operational during business hours but ran 24/7 due to stagnant development and staging.

3:Orphaned resources ($156,000): Load balancers served zero traffic because of how 89 instances were being tied to depreciated projects.

4:Inefficient Databases ($104,000]: Opportunities were being allocated to make use of autoscaling with a total average of 30% in utilisation,but the databases were provisioned only for peak capacity and static allocation.

5:Redundant backups ($52,00): Overlapped backup configurations created alot of irrelevant duplicates.

6:Obsolete instances ($28,000): While resources ran on past generational instances,newer types were offered 40% which boosted it's performance.

7:Unoptimised storage ($8,000): Data were frequently being accessed on premium SSD servers instead of the cost-effective tiers.

Total savings identified: $847,000 annually.

“I thought $100,000 was being wasted annually,” Park admits.”$847,000 was indeed both shocking and embarrassing.This money could have been used to hire alot of skilled expertise.”

The Solution:Automated Optimisation 

CloudOptimise didn't just identify problems but provided one-click implementation for alot of fixes such as;

Automated scheduling:Facilities with no sign of production were shut down automatically.

Saved $187,000 annually.

Right sized recommendations:CloudOptimise suggested specific type of instances for each of the overprovisioned resource. Vertex approved the request and this helped them save $312,000.

Orphan cleanup:Adoption of the platform led to the total termination of unused resources.Vertex followed up with alot of erasures over the next two weeks. Saved $156,000.

Database optimisation:In order to reduce baseline costs during operations,CloudOptimise recommended using autoscaling configurations because of the need to maintain performance during traffic spikes.

Saved $104,000.

Backup consolidation:The identification of overlapped backup jobs by CloudOptimise lead to a proposal for a streamlined strategy which aligned perfectly well with Vertex disaster recovery requirements.

Saved $28,000.

Implementation timeline:6 weeks from first scan to optimisation.

The Results:$847,000 Saved With Zero Downtime 

Six months after it's implementation,the following results were achieved;

•43% decrease in both cloud expenses and cost reduction.

•$847,000 saved annually.

•Performance impact made customer metrics (latency,uptime,error rates) stay unchanged.

•Steady connections being experienced in both time and value were due to the actionable recommendations. 

•Payback period for up to 2-3 months.

•Consistent value due to CloudOptimise close monitoring of new optimisation opportunities.

“Not a single customer was affected after our cloud bill was cut into half,” Park says.

“That money turned disastrous year into a profitable one.”

CFO Jennifer Morrison quantifies the whole business impact,saying”The $847,000 being saved annually was used to cover our entire product development budget for Q1.Two major updates were launched unlike in the past where it would have just evaporated into AWS.”

Secondary Benefits 

The financial impact was less obvious but equally valuable due to operational improvements.

Visibility:Park now had enough data to carry out decisions because Vertex was now fully aware of where it's money went to.CloudOptimise dashboard showed variations cost of each team,project and features.

Accountability:With clear cost attribution,engineering teams became mindful of infrastructure inefficiency.”Developers started asking ‘do we really need this instance type?’before any provisions for resources were made,” Park notes.

Capacity Planning:Vertex could forecast future spending based on it's customer growth projections unlike the historical cost trends which revealed only patterns.

Faster incident responses:When costs spiked so unexpectedly,Park's team was alerted by CloudOptimise within minutes.Instead of swooping into investigating thousands of line items,the anomalies were traced to a cache of specific resources.

“The cost savings being paid for CloudOptimise really paid off,”Park reflects.”But the operational clarity might turn out to be a long-term commitment.”

Lessons For Other Companies 

Vertex's experience offers guidance for similar companies wrestling with rising cost of cloud infrastructure

1:Wastes are accumulated invisibly:No one intends to overspend but infrastructural complexity creates alot of blindspots.Due to the occasional inefficiency experienced in audits,close monitoring systems are set up to catch waste before it emerges.

2:Manual optimisation doesn't scale. AI handling of workload effortlessly can't be compared to Vertex infrastructural team of six which couldn't even manually analyse 2,400 instances.

3:Optimisation requires no performance tradeoffs.Vertex cut it's costs by 43% without any of it's services been degraded.The myth that says “cheaper means slower” doesn't apply because waste elimination reduces alot of costs.

4:Speed matters alot.Months of intense audits cost Vertex $11,000 weekly.

5:Intelligent optimisation systems were being setup to target specific waste like blanket reductions whilst preserving what actually works.

The Road Ahead 

Vertex still uses CloudOptimise for it's ongoing optimisation. As more customers and new features are being added,the platform maintains it's grip through it's efficient opportunities.

Park's team now provisions infrastructure differently.Before new services are launched, references are made to CloudOptimise historical data of selected instance types. The whole architectural reviews includes cost considerations alongside efficient performance and reliability.

“CloudOptimise has changed our perspective on cloud infrastructure,”Park concludes.”We used to ask ‘will this work?’ Now we ask ‘will this work efficiently?That shift in mindset really compounds over time.”

For Vertex,$847,000 saved annually represents more than just reclaimed expenses. It represents sustainability which is the ability to grow profitable instead of trading away it's revenue.

About CloudOptimise: CloudOptimise uses artificial intelligence to analyse cloud infrastructure and identify cost optimisation opportunities.Companies using CloudOptimise have drastically reduced their annual cloud budget by 38%.