COST05-BP06 - Perform cost analysis for different usage over time

Implementation guidance

When selecting services, analyze cost not just for current usage but across the expected range of usage over time. A service that is cheapest at today’s volume may not be the most cost-effective as the workload grows, shrinks, or changes its access pattern. Modeling cost across future usage scenarios prevents selecting a service that becomes expensive at scale.

Model usage over time

Project usage scenarios: Define low, expected, and high usage scenarios over a meaningful time horizon, based on business forecasts and historical growth.

Compare candidate services across scenarios: For each candidate service or configuration, estimate cost under each usage scenario rather than at a single point. Include the effect of pricing models (on-demand, Savings Plans, Reserved Instances, tiered pricing) that change unit cost as volume changes.

Account for the full cost curve: Consider how cost scales — linearly, in steps, or with volume discounts — and identify break-even points where a different service or pricing model becomes cheaper.

Use the analysis to decide

Select for the expected lifecycle: Choose the service that is most cost-effective across the realistic range of future usage, not only at launch.

Re-evaluate as usage changes: Revisit the analysis when actual usage diverges from projections, and use it together with COST05’s other best practices to keep service selection cost-effective over time.

Automate the modeling: Use the AWS Pricing Calculator and your detailed CUR data (queried in Amazon Athena — the AWS analytics service, not an internal agent) to make the analysis repeatable.

AWS Services to Consider

Related Resources