The Cloud IS Expensive. Here’s What to Actually Do About It Before Your CFO Notices.

Mike Julian at Duckbill Group wrote a piece a few months back that’s been sitting in my head. The thesis: FinOps is solving the wrong problem. The cloud isn’t expensive, he argues. The real problem is predictability. Business leadership doesn’t care if the bill is $1M or $100M. They care whether they can forecast it, explain it, and control it.

He’s got a point. AWS has 2.3 million SKUs. Consumption-based billing means your infrastructure cost is a function of customer behavior, engineering decisions, and market conditions. That’s genuinely hard to model. And he’s right that FinOps teams spend most of their time on budgets, forecasts, vendor negotiations, and stakeholder reporting, not hunting down idle instances.

But there’s a gap in the argument, and it matters for anyone not spending nine figures on AWS.

Predictability Starts With Fixing the Basics

Mike’s framing works for the customer base Duckbill serves. Skyway is explicitly “cloud cost management for the nine-figure club.” When your bill is $100M, a 30% swing is $30M and the priority is absolutely getting a handle on what drove it, building forecasts, and structuring your contracts.

But most companies aren’t in the nine-figure club. We work with companies spending $250K to $250M a year on AWS. For them, the bill isn’t unpredictable because of contract complexity or multi-vendor data normalization. It’s unpredictable because nobody did the work.

Here’s what a 30% spike actually looks like when you open Cost Explorer and sort by service:

  • A team spun up 50 instances for a project that finished, but nobody shut them down
  • EBS volumes accumulated for months because nobody enabled gp3 migration
  • Someone moved a database to a larger instance “temporarily” and forgot to move it back
  • S3 storage grew unchecked because no lifecycle policies were set up

These aren’t mysteries. They’re not unpredictable. They’re a to-do list that never got done. And the CFO asking “why did our bill jump 30%” deserves a better answer than “let me spin up a data platform to investigate.” The answer is sitting right there in Cost Explorer. What’s missing is someone to fix the things that cause the spikes.

The Real Gap is Execution, Not Visibility

Mike is right (again) that the industry built tools for finding savings and left practitioners to figure out the other 90% on their own. Where I disagree is the conclusion. He argues the answer is a data platform that makes consumption explorable and predictable across the organization.

But every time CloudFix connects to a new AWS account, whether it’s $250K or $250M, we find the same things:

  • gp2 volumes that should have been migrated years ago
  • S3 buckets with no lifecycle policies
  • On-demand instances where Savings Plans or Reserved Instances would cut costs 30-72%
  • Idle resources nobody turned off

The problems repeat. The fixes repeat. The only variable is whether anyone implemented them.

You can have the best forecasting model in the world. If nobody migrates those gp2 volumes, your forecast will still be wrong. Predictability doesn’t come from better reporting. It comes from fixing the things that make costs unpredictable in the first place.

Run a free CloudFix assessment and see exactly what’s driving your bill. Five minutes to connect. 24 hours to get a report with dollar amounts next to every fix. CloudFix implements the fixes. You approve what you want, skip the rest. Add RightSpend for up to 55% off EC2 with zero long-term commitments.

Two Problems, Two Tools

To be clear: I’m not saying Mike is wrong. Contract intelligence, private pricing management, and cross-vendor cost modeling are real problems. If you’re spending nine figures across AWS, Databricks, Snowflake, and a dozen other vendors, you need what Skyway is building. Duckbill knows that world better than anyone.

But for the vast majority of AWS customers, the path to a predictable bill runs through execution, not analysis. Fix the gp2 volumes. Set up the lifecycle policies. Purchase the Savings Plans. Automate the fixes so they keep running.

Then your bill stops spiking. And you don’t need a data platform to tell you why.

CloudFix finds and implements AWS optimizations automatically across EC2, EBS, S3, RDS, and Lambda. RightSpend delivers up to 55% EC2 savings with zero commitments. Both available on the AWS Marketplace. See what you’d save.