AWS Cost Explorer told us EC2 was up $2K. We still couldn't answer which team.

Finance sends an email. "AWS bill is up $4K from last month. Can engineering explain?"

Your CTO opens Cost Explorer. Pulls up last month vs this month. EC2 is up $1,800. RDS up $900. Data transfer up $600.

She takes this to the all-hands. "EC2 was up $1,800. Which team? Which project?"

Nobody can answer quickly.

This conversation happens at almost every engineering-led startup, every month.

What Cost Explorer actually shows you

AWS Cost Explorer is a genuinely good tool for what it does. It shows you:

• Total spend by service (EC2, RDS, S3...)
• Spend over time (daily, monthly, yearly)
• Usage type breakdown
• Reserved Instance and Savings Plan coverage

What it doesn't show you:

• Which specific resources are idle vs active
• Which team or project owns a resource
• Which stopped instance has been running up charges for 60 days
• Which EBS volumes are orphaned and not attached to anything
• What's safe to delete and what isn't

Cost Explorer answers "how much." It doesn't answer "why" or "what should we do about it."

The ownership problem

Here's the scenario that plays out in most startups with 3+ engineers:

• Team A spins up an EC2 instance for a new feature
• Feature gets deprioritized
• Instance gets stopped "temporarily"
• Six months later, nobody remembers who owns it
• It's still billing $14/month in EBS storage

Cost Explorer will show you EC2 spend is up. It won't tell you that instance belongs to Team A, was stopped in November, has zero network activity since October, and could be safely deleted after a 5-minute review.

That context gap is the real problem.

Why tagging doesn't fully solve this

Most teams know they should tag resources with team, environment, project. Most teams don't do it consistently.

Why? Because tagging at creation time is easy to forget, and retroactively tagging hundreds of resources is painful manual work.

Even teams with good tagging discipline have:

• Resources created before the tagging policy existed
• Third-party resources that don't support tagging
• Auto-scaling instances that don't inherit the right tags

Cost Explorer shows you tag gaps as a cost bucket labeled "no tag." It can't tell you who probably owns the untagged resource based on creation time, region, naming pattern, and VPC membership.

What InfraDesk adds on top of Cost Explorer

InfraDesk is not a replacement for Cost Explorer — it works alongside it.

Cost Explorer tells you EC2 is up $1,800. InfraDesk tells you:

• 3 stopped instances have been billing for 45–90 days: $51/month combined
• 2 unattached EBS volumes from deleted instances: $18/month
• 1 unused Elastic IP: $3.60/month
• jenkins-setup — last activity 61 days ago, no CloudFormation membership, no owner tag: $17/month, likely safe to review

Now when finance asks "why is EC2 up $1,800?" you can answer specifically. And more importantly, you can action specific items with evidence — not guesswork.

The conversation that changes

Before InfraDesk, the monthly finance-engineering conversation sounds like:

"EC2 is up. We think it might be the new microservices rollout. Or maybe the data pipeline. We're looking into it."

After getting visibility into waste:

"EC2 is up $1,800. $89 of that is identifiable waste — 3 stopped instances and 2 orphan EBS volumes we can clean up this week. The remaining $1,711 is real usage increase from the new microservices rollout, which is expected."

The second answer builds trust. The first one doesn't.

Start with visibility

You already have Cost Explorer. It's free and useful. Keep using it for spend trends and Reserved Instance planning.

What you probably don't have yet: a clear view of which specific resources are waste, who owns them, and whether it's safe to act.

That's the gap. And it's solvable.