AWS Cost Optimization: 5 Things Utah Businesses Are Overpaying For

AWS is a remarkable platform. It's also remarkably good at extracting money from your business if nobody's paying attention.

I've seen it over and over with Utah businesses — someone set up the infrastructure a year or two ago, it works fine, and nobody's looked at the bill since. Then one month the invoice is 40% higher than expected and everyone's scrambling to figure out why.

The thing is, AWS doesn't optimize itself. It's designed to make spinning up resources easy. Spinning them down? That's on you. And if you're not actively managing your spend, you're almost certainly leaving money on the table.

Here are the five places I see businesses overpaying most consistently.

1. Over-provisioned EC2 instances

This is the big one. It's also the easiest to fix.

When someone first sets up your infrastructure, they tend to pick instance sizes with a comfortable margin. Makes sense — nobody wants to under-provision and have performance issues. But "comfortable margin" often means you're running instances that are three or four times bigger than what your workload actually needs.

A t3.xlarge running at 8% average CPU utilization doesn't need to be a t3.xlarge. That's like renting a school bus to commute to work alone.

AWS has tools for this — Compute Optimizer, CloudWatch metrics — that will tell you exactly how much of your provisioned capacity you're actually using. Right-sizing is usually the single biggest source of savings. We regularly see businesses cut 30-50% off their compute costs just by matching instance sizes to actual workload.

And it's not a one-time thing. Workloads change. That instance that was perfectly sized six months ago might be way too big now. Or too small. You need to check regularly.

2. Forgotten resources

This one's insidious because each individual item looks small. But they add up fast.

Old EBS snapshots from instances you deleted months ago? Still being billed. That Elastic IP you allocated for a test server and never released? $3.65 a month — doesn't sound like much until you have fifteen of them. Orphaned load balancers sitting in front of nothing? Still running the meter.

I did a cleanup for a client last year and found over $400/month in resources that were attached to absolutely nothing. Nobody knew they were there. The infrastructure had evolved over time and these things just got left behind like furniture in a storage unit you forgot about.

The fix is boring but effective: regular resource audits. Go through your account, tag everything, and delete what's not being used. AWS Trusted Advisor can flag some of this automatically, but it doesn't catch everything.

3. Data transfer costs

This is the one that blindsides people.

Data transfer into AWS is free. Data transfer out — egress — is not. And the pricing isn't exactly intuitive. Moving data between availability zones? That costs money. Between regions? More money. Out to the internet? Even more.

Most people don't think about this when they're designing their architecture. Then they build an application that shuffles data between services in different AZs, or serves large files directly from S3 to users, and suddenly data transfer is 20% of their bill.

Some fixes: use CloudFront for content delivery instead of serving directly from S3 or EC2. Keep services that talk to each other frequently in the same AZ. Use VPC endpoints for AWS service traffic instead of routing through the internet. Compress data before transferring it.

None of this is hard. But you have to know to look for it.

4. No Reserved Instances or Savings Plans

If you're running workloads on AWS that you know will be there in a year — and most businesses are — paying on-demand pricing is like paying rack rate at a hotel every single night. You're leaving 30-60% savings on the table.

Reserved Instances and Savings Plans are AWS's way of giving you a discount for committing to a certain level of usage. The commitment can be one or three years. You can pay all upfront, partially upfront, or nothing upfront (though the discount is smaller).

The math is straightforward. If you've been running the same set of EC2 instances for six months and they're not going anywhere, convert them to Reserved Instances. A one-year no-upfront reservation typically saves 30-40%. Three-year all-upfront can save over 60%.

Savings Plans are more flexible — they apply to usage across instance families and even across services like Fargate and Lambda. For most businesses, they're the easier option.

The hesitation I hear is "what if our needs change?" Fair concern. But even a conservative commitment on your baseline usage — the stuff you know you'll need no matter what — saves real money. You can always run the variable stuff on-demand.

5. No billing alerts or budget caps

This is less about optimization and more about not flying blind.

I've talked to business owners who didn't realize their AWS bill had doubled until they saw the charge on their credit card statement. By then, the damage is done. You've already paid for a month of waste.

AWS Budgets lets you set spending thresholds and get alerts when you're approaching them. It takes about ten minutes to set up. You can set alerts at 50%, 80%, and 100% of your expected spend. You can break it down by service, by account, by tag.

There's also AWS Cost Anomaly Detection, which uses machine learning to flag unusual spending patterns. If someone accidentally spins up a fleet of GPU instances or a misconfigured service starts generating massive data transfer, you'll know about it the same day instead of the end of the month.

This is table stakes. If you don't have billing alerts set up, stop reading this and go do it right now. Seriously.

What a managed AWS review looks like

At Wolfgang, we do monthly AWS cost reviews as part of our managed services retainer. Here's what that actually involves:

We spend one to two hours each month going through your account. We pull utilization data, check for unused resources, review your Reserved Instance coverage, and look at data transfer patterns. We run the numbers on right-sizing opportunities and flag anything that looks off.

We set up automated alerts so nothing sneaks past between reviews. And we give you a clear summary of what we found, what we recommend, and how much it'll save.

It's not glamorous work. But I've seen it save clients anywhere from $500 to $5,000 a month depending on their AWS footprint. Over a year, that's real money.

Start saving

If you haven't reviewed your AWS costs in the last 90 days, you're almost certainly overpaying. The good news is that most of the fixes are straightforward — they just require someone to actually sit down and do the work.

We do monthly AWS reviews as part of our managed services retainer. It's one of those things that pays for itself almost immediately.

Want to talk about your AWS setup? Reach out and we'll take a look. No pressure, just an honest assessment of where you stand and where you can save.