Choosing an AWS cost management strategy

How you manage the costs of your cloud investment depends on what you’re trying to do, as well as how you’re tracking performance. If your organization is in a growth phase, then your KPIs may focus on market share, customer growth, and getting the best return on investment. If your focus is on reducing operating costs, then you will concentrate more on achieving the right balance between spending and delivering for your customers.

Set your budget based on the trend shown by your KPIs, as well as your available funds. Then set the budget thresholds using AWS Budgets for cost or resource usage and to set utilization or coverage targets for your Reserved Instances (RIs) and Savings Plans.

Monitor how the budget portfolio is progressing in your AWS Budgets Reports dashboard. Create notification alerts, which will be sent whenever cost and usage either reach your forecasts or exceed your threshold limits.

One key strategy for controlling cost is to have control over which services can be used by which members of your organization. If you need this kind of access control, use IAM to control individual and group access to your resources. This involves being able to create, manage, and grant permissions to IAM users to let them access the appropriate resources for their role and location. Use AWS Organizations to enable automatic policy-based account creation, management, and billing at scale.

It’s useful to get alerts that let you know if you’ve exceeded your budget, but it’s even more useful to see what is causing your bill to exceed your budget. Use AWS Cost Explorer, to see which AWS services are increasing your spend.

To see which parts of your organization drive increased spend and usage, you can categorize your resources and spending using Cost Allocation Tags with AWS Cost Explorer.

If you want your cost management strategy to scale with your usage, it is important to get detailed reporting when you have multiple users working with multiple AWS services.

You can use the billing dashboard on the AWS Billing Console to:

It’s not only about the numbers. Your cost optimization strategy must also align with the logic of your business. AWS Billing Conductor lets you customize your AWS pricing and billing report to provide that alignment with your business logic. To help address your cost allocation needs, group accounts according to their financial relationships, adjust ratings, share credits and savings, or add overhead costs.

With AWS Billing Conductor, you can customize a second, alternative version of your monthly billing data. This alternative version models the billing relationship between you and your customers or business units, but doesn't change the way that AWS bills you. Using the alternative version provides you with a mechanism to configure, generate, and display rates to certain customers over a given billing period. Analyze the difference between the rates that you apply to your groupings, relative to the actual rates for those same accounts from AWS.

Customers with predictable, steady workloads on Amazon EC2, Amazon SageMaker, and Amazon RDS should consider Reserved Instances (RIs) to save up to 75% over equivalent on-demand capacity. RIs are available in three options: all upfront, partial upfront, and no upfront. To maximize your savings, when you buy RIs, the larger the upfront payment, the greater the discount. Partial upfront RIs offer lower discounts, but give you the option to spend less up front. Lastly, you can choose to spend nothing up front and receive a smaller discount, which lets you free up capital to spend on other projects.

We offer Reserved Instance and Savings Plan purchase recommendations through AWS Cost Explorer, based on your past usage. AWS Cost Explorer identifies and recommends the commitment value that it estimates will result in the largest savings.

Another key element of cost management is identifying and removing any idle or over-provisioned resources. AWS Cost Explorer resource recommendations let you view top-level KPIs for rightsizing and instance selection recommendations.

These recommendations identify idle and underutilized instances, and search for smaller instance sizes in the same instance family.