AWS Cost Optimization: How Caribbean Enterprises Waste 25-40% of Their Cloud Spend

We have never audited a Caribbean enterprise AWS account and NOT found at least 25% waste. The range is typically 25-40% of total monthly spend that can be eliminated without any impact on performance or availability. That is not a criticism of the teams managing these accounts — it is a structural problem with how cloud resources are provisioned and forgotten.

The five most common waste patterns we find in every audit:

Pattern 1: Oversized instances. The default behaviour when provisioning an EC2 instance or RDS database is to choose something larger than needed "just in case." That t3.xlarge running your staging environment at 3% CPU utilisation? It should be a t3.small. That db.r5.large hosting a database with 2GB of data? A db.t3.medium would perform identically. We typically find 30-50% of instances are at least one size too large. AWS Compute Optimizer will tell you this for free — but someone has to look at it.

Pattern 2: No Reserved Instances or Savings Plans. If you have workloads that run 24/7 (and most production systems do), you are overpaying by 30-40% on those resources if you are using On-Demand pricing. A 1-year Compute Savings Plan with no upfront payment saves 20%. A 3-year plan saves 36%. For a Caribbean enterprise spending $5,000 per month on EC2, that is $1,500-$1,800 per month in savings for a commitment you were going to make anyway. Yet we consistently find companies running entirely on On-Demand after 2-3 years in production.

Pattern 3: Zombie resources. These are resources provisioned for a project that ended, a test that completed, or a developer who left. NAT Gateways processing zero traffic: $33/month each. Elastic IPs attached to nothing: $3.60/month each. EBS volumes with no attached instance: $8-80/month each. RDS instances nobody queries: $15-200/month each. Individually small, collectively devastating. We recently found $111/month in zombie resources on a single account in an 8-minute audit.

Pattern 4: Unoptimised storage. S3 buckets with no lifecycle policies accumulate data forever. That 500GB of CloudWatch logs from 2023? Still in Standard storage at $0.023/GB/month instead of Glacier at $0.004/GB. Incomplete multipart uploads consuming space. EBS snapshots from deleted volumes. Old AMIs nobody uses. Storage waste is insidious because each item is cheap but they accumulate relentlessly.

Pattern 5: Non-production environments running 24/7. Your development and staging environments probably do not need to run at 2 AM on Sunday. Yet they do — because nobody set up a schedule to stop them. An Auto Scaling schedule or a simple Lambda that stops dev instances at 7 PM and starts them at 7 AM saves 60% on those resources immediately. For a team with 3 dev environments running t3.large instances with RDS, that is $300-500 per month saved overnight.

The FinOps basics for a 10-person IT team do not require a dedicated FinOps engineer or expensive tooling. Start with these four practices: First, enable AWS Cost Explorer and check it weekly (not monthly). The weekly cadence catches anomalies before they compound. Second, set AWS Budgets with alerts at 80% and 100% of expected spend. This costs nothing and prevents surprises. Third, tag everything. If you cannot tell which project, team, or environment a resource belongs to, you cannot optimise it. Fourth, run Trusted Advisor monthly. The cost optimisation checks are free with Business Support and surface the low-hanging fruit automatically.

Tools you can use today: AWS Compute Optimizer (free — analyses your instance utilisation and recommends right-sizing). AWS Cost Explorer (included — visualise spend trends, forecast future costs). AWS Trusted Advisor (Business Support — identifies idle resources, underutilised instances, missing reservations). AWS Instance Scheduler (free solution — automates start/stop of non-production resources). For more sophisticated analysis, tools like CloudHealth, Spot.io, or nOps provide deeper recommendations, but the native AWS tools cover 80% of the opportunity at zero additional cost.

The audit process we run for clients takes 4-6 hours and typically identifies $500-$3,000 per month in immediate savings. The structure: inventory all resources by service and region, identify resources with zero or near-zero utilisation over 14 days, flag On-Demand workloads running longer than 6 months (candidates for Savings Plans), check for missing lifecycle policies on S3 and CloudWatch, verify that non-production environments have schedules, and compare instance sizes against Compute Optimizer recommendations.

The real question is not whether you have waste — you do. The question is whether you have a process to catch it continuously. Cloud costs are not a one-time optimisation. They require ongoing attention because teams provision new resources weekly and rarely deprovision them. A monthly 30-minute review of Cost Explorer anomalies prevents most waste from accumulating.

If you are a Caribbean enterprise spending more than $3,000 per month on AWS and have not done a formal cost review in the last 6 months, you are almost certainly overpaying by $750-$1,200 per month. That is $9,000-$14,400 per year — enough to fund a significant infrastructure improvement or hire a part-time contractor. The savings are there. Someone just needs to look.