Why is my AWS bill so high, and how do I cut it?
The usual reasons a small D2C brand's AWS bill creeps up, how to find where the money goes, and the safe steps to cut it without hurting sale-day reliability.
Anadoris Digital · · 4 min read

Few things worry a founder more than an AWS bill that keeps rising while nobody can explain why. Usage-based pricing is flexible, but it also means that every forgotten server, oversized database and verbose log quietly adds to the total. Bills rarely jump because of one big decision; they creep because of dozens of small ones nobody revisited.
The good news is that most small brands can bring their bill under control without touching anything customers rely on. The work is mostly housekeeping: seeing where money goes, removing waste and matching resources to real usage. It is worth doing before your next big sale, so that the extra capacity you add for the event is not stacked on top of waste.
Find out where the money actually goes
Open AWS Cost Explorer and group costs by service, then by region, then by usage type. This usually reveals the answer quickly: one database, a group of servers, data transfer, or a service you did not realise you were using. Look at the trend over several months, not just the latest bill, and note when increases began. Often the start date lines up with a launch, a new app or a developer experiment.
- Group by service to see which parts of AWS cost the most.
- Group by region to spot resources running somewhere you do not use.
- Group by usage type to separate compute, storage, data transfer and requests.
- Check for services you do not recognise, which often point to forgotten experiments.
- Compare months to find when growth began and what changed around then.
The usual suspects
- Servers and databases running at a fraction of their capacity, sized for a peak that rarely happens.
- Staging, test and demo environments left running around the clock.
- Unattached storage volumes, old snapshots and unused elastic IP addresses.
- Log groups with no retention limit and debug logging left on in production.
- NAT gateways and data transfer between zones, regions or out to the internet.
- S3 buckets full of old exports, backups and duplicates in expensive storage classes.
- Resources in regions you never meant to use.
Data transfer deserves a special mention because it is easy to miss. Traffic leaving AWS, moving between regions, or passing through a NAT gateway is charged, and an app that downloads large files repeatedly or a database in a different zone from its app can run up costs quietly. Check current AWS pricing for these charges and keep apps and their data close together.
Safe ways to cut costs
Start with the changes that carry no risk: deleting resources nobody uses and setting limits on logs and storage. Then move to changes that need a little testing, such as resizing servers. Leave commitments for last, once usage is clean and predictable.
- Delete clearly unused resources after confirming with their owners and taking a final backup if needed.
- Schedule staging and test environments to switch off outside working hours.
- Right-size servers and databases using AWS Compute Optimizer recommendations and real usage data.
- Move to newer, more efficient instance types, including Graviton-based options where your software supports them.
- Set retention on log groups and lifecycle rules on S3 buckets to move or delete old data.
- Cache more at the CDN so fewer requests reach your servers.
Savings Plans and Reserved Instances
If part of your usage is steady, such as a production database or servers that run all year, AWS offers lower rates in exchange for a commitment over a fixed term. Savings Plans apply to compute usage across services and instance types; Reserved Instances apply to specific services such as databases. They can reduce costs noticeably, but you pay for the commitment whether or not you use it.
For a small brand, commit only to the baseline you are confident will run for the whole term, after cleaning up waste and right-sizing. Never commit based on a sale-day peak. Check current AWS pricing and terms before buying, and review coverage every few months.
Make costs visible so they stay down
Clean-ups wear off if nobody watches. Tag every resource with its project, environment and owner, and activate those tags for cost allocation so Cost Explorer can group by them. Set budgets with alerts, turn on AWS Cost Anomaly Detection to flag unusual spending, and hold a short monthly review where each owner explains their line.
Where to start
This week, open Cost Explorer, list your five most expensive services, and for each one write down which resources drive the cost and who owns them. Delete what nobody claims, set log retention everywhere, and schedule staging to switch off at night. Then set a monthly budget with alerts so the next increase reaches you before the invoice does.
If you would like help with a structured cost review, Anadoris Digital includes one in its AWS and cloud management work.
Related service: AWS & Cloud Management · More on this topic: AWS & Cloud Management articles
