Cloud cost optimization: A strategic issue for Canadian businesses

As cloud spend accounts for a growing share of IT budgets, AWS cost optimization has become a strategic priority. According to the 2024 State of FinOps report, 82% of organizations consider cloud cost management a major challenge, and 78% believe that more than 20% of their cloud spend is avoidable waste.

AWS offers two main mechanisms to reduce infrastructure costs through longer-term commitment agreements: Reserved Instances and Savings Plans. Although both options offer substantial discounts compared to on-demand pricing, they work differently and are suited to different situations. Choosing the wrong strategy can lead to thousands of dollars in unnecessary spending or, conversely, a costly lack of operational flexibility.

In this article, we explore the differences between these two approaches and present concrete commitment strategies, from short-term flexibility to long-term optimization, to help you maximize savings while maintaining the agility your growth requires.

Understanding the options: Reserved Instances vs Savings Plans

Reserved Instances (RI): the specific commitment

Reserved Instances are AWS’s historical mechanism for offering discounts in exchange for a commitment. They work on a simple model: you commit to using a specific instance, including type, region and operating system, for a set period, and AWS gives you a substantial discount.

Key RI characteristics:

• Commitment to a specific instance type, such as m5.large
• Discounts of up to 72%, depending on the commitment
• Less flexibility: changing the instance type requires an exchange process
• Ideal for highly stable and predictable workloads

Savings Plans: modern flexibility

Introduced in 2019, Savings Plans represent the modern evolution of AWS cost optimization. Unlike RI, you commit to an hourly spend amount, such as $10/hour, rather than specific instances.

Key Savings Plans characteristics:

• Commitment to a spend amount, not to specific instances
• Full flexibility across instance types, regions and instance families
• Automatically apply to EC2, Lambda and Fargate
• Discounts of up to 72%, similar to RI, but with much greater flexibility

Practical verdict: For most modern organizations practicing DevOps and architectural agility, Savings Plans offer the best balance between savings and operational flexibility. Reserved Instances remain relevant for very specific cases where the workload is perfectly predictable and stable over several years.

Commitment strategies: short term vs long term

AWS offers several Savings Plans commitment terms, each with a different level of savings. The choice between these options depends on your ability to forecast future consumption and your tolerance for the risk of overcommitting.

1-year commitment – No upfront payment

Savings: 10% per year on compute costs

This strategy is the least financially binding option. It is especially suited to organizations with variable or hard-to-predict workloads, as well as those that do not yet have a clear view of their growth trajectory.

Typical use cases:

• Startups in the product-market validation phase
• Companies testing new cloud architectures
• Organizations gradually migrating to AWS
• Development and test environments with fluctuating usage

Concrete example: A B2B SaaS startup that has just secured its Series A funding and anticipates rapid but uncertain growth may choose this strategy. The 10% savings already provide meaningful budget optimization without locking the company in for three years.

3-year commitment – No upfront payment

Savings: 33% per year on compute costs

This strategy delivers the highest savings, but involves a long-term commitment. It is recommended for organizations with stable and predictable workloads, with little expected variation over three years.

Typical use cases:

• Companies with production workloads that have been stable for several years
• Legacy applications with predictable consumption
• Regulated workloads requiring constant infrastructure
• Organizations looking for maximum budget predictability

Watch for traps: A 33% savings commitment may look attractive, but it creates risk if your architecture changes significantly. A migration to containers, such as ECS or EKS, increased adoption of Lambda, or instance consolidation can make that commitment less relevant.

The progressive 1/3 – 2/3 strategy: controlled growth

This approach allows you to gradually increase your commitment over three years, offering maximum flexibility to adjust your Savings Plans based on the real growth of your infrastructure.

How the staggering works

This approach ensures that approximately one-third of your Savings Plans expire each year. Instead of being stuck with a massive commitment expiring all at once, you can progressively readjust your commitments based on the actual evolution of your consumption.

Commitment breakdown:

Year	3-year commitment	1-year commitment	Estimated savings
1	1/3 of the recommended amount	2/3 of the recommended amount	17.67%
2	2/3 of the recommended amount	1/3 of the recommended amount	25.33%
3	100% of the recommended amount	—	33%

 

Key advantages of this strategy

Staggered renewals: Each year, approximately one-third of your commitments expire, allowing you to readjust your strategy based on real observed growth rather than anticipated growth.

Risk reduction: By avoiding a massive commitment in the first year, you limit the risk of overcommitting, which means paying for unused capacity, or undercommitting, which means missing potential savings.

Architectural flexibility: This approach gives you the freedom to evolve your cloud architecture, whether that means migrating to serverless services, adopting containers, or consolidating instances, without being constrained by rigid commitments.

Ideal use case: Organizations in a growth phase, such as funded scale-ups or expanding SMBs, that want to maintain significant room for adjustment during the first two years. This strategy is especially relevant for companies that expect growth but are still uncertain about the exact pace.

The balanced 1/2 – 1/2 strategy: The smart compromise

This approach offers a compromise between flexibility and savings optimization starting in the first year. It is well suited to organizations with better visibility into their future consumption.

How the staggering works

Like the progressive strategy, this approach creates a staggered structure where half of your commitment expires after one year, allowing you to readjust your strategy midway before the final expiry.

Commitment breakdown:

Year	3-year commitment	1-year commitment	Estimated savings
1	1/2 of the recommended amount	1/2 of the recommended amount	21.45%
2	100% of the recommended amount	—	33%

 

Key advantages of this strategy

Midpoint readjustment: After the first year, when 50% of your commitments expire, you have concrete data on your real consumption to optimize your strategy for year two.

Optimized savings from year one: With 21.45% savings in the first year, compared to 17.67% for the progressive strategy, this approach delivers a faster return on investment while maintaining flexibility.

Accelerated transition: Starting in year two, you reach the maximum savings level of 33%, enabling full budget optimization more quickly.

Ideal use case: Organizations anticipating moderate growth, whose workload has stabilized in recent years and that want to maximize savings while maintaining some initial flexibility. This strategy is especially relevant for established SMBs and mid-market companies with a stable customer base.

Conclusion: Choosing the strategy that fits your reality

The choice between Reserved Instances and Savings Plans, as well as the commitment strategy to prioritize, fundamentally depends on three factors: your ability to forecast future consumption, your tolerance for financial risk, and the maturity of your cloud architecture.

Savings Plans are now the preferred option for most modern organizations, offering the optimal balance between substantial savings, up to 33%, and architectural flexibility. For growing organizations, the progressive 1/3 – 2/3 strategy provides an important safety net, while the balanced 1/2 – 1/2 strategy is better suited to more stable environments looking to maximize savings quickly.

AWS cost optimization is not a one-time decision. It is an ongoing process that requires regular analysis of your real consumption and a strong understanding of AWS pricing mechanisms. The wrong strategy can result in tens of thousands of dollars in unnecessary spend or missed savings opportunities.

At Unicorne, we help Quebec businesses analyze their AWS consumption and implement tailored optimization strategies. Our data-driven approach and deep expertise in AWS pricing mechanisms allow us to identify savings opportunities specific to each organization. Contact us for an audit of your AWS infrastructure and find out how to maximize your savings while maintaining the agility your growth requires.

FAQs

What is the difference between a Savings Plan and a reservation?

The main difference is flexibility. A reservation, often called a Reserved Instance, is tied to specific infrastructure choices such as instance type, region, operating system and term. A Savings Plan is based on a committed hourly spend instead. That means it can apply more flexibly across eligible AWS compute usage, which makes it better suited to organizations whose workloads, architectures or growth plans may change.

When should you consider Reserved Instances or Savings Plans instead of On-Demand pricing?

Reserved Instances or Savings Plans are worth considering when you have AWS workloads that run consistently and are unlikely to disappear in the short term. On-Demand pricing is useful for temporary, unpredictable or experimental workloads, but it is usually more expensive over time. If part of your AWS usage is steady, committing to that baseline can help reduce costs without covering every workload.

Do Savings Plans apply to RDS?

Standard AWS Savings Plans do not apply to Amazon RDS. They mainly apply to eligible compute usage such as EC2, AWS Fargate and AWS Lambda. For RDS, AWS offers Reserved Instances, which can reduce database costs when usage is stable and predictable. This is why a good AWS cost optimization strategy often looks at compute and database commitments separately.

What are the advantages of Reserved Instances?

Reserved Instances can offer significant savings compared to On-Demand pricing when workloads are stable, predictable and unlikely to change. They can also help improve budget planning because costs become easier to forecast over the commitment period. However, they are less flexible than Savings Plans, so they are best suited to mature environments where the infrastructure footprint is well understood.

Are AWS Savings Plans better than Reserved Instances?

Not always. Savings Plans are often better for modern cloud environments because they offer more flexibility across eligible compute usage. Reserved Instances may still make sense for very stable workloads, especially when specific services such as RDS are involved. The best option depends on how predictable your usage is, how often your architecture changes and how much flexibility your team needs.

How do you avoid overcommitting with AWS Savings Plans or Reserved Instances?

The safest way to avoid overcommitting is to base commitments on your stable baseline usage, not on peak traffic or aggressive growth forecasts. Before committing, review historical usage, upcoming architecture changes, migration plans and product roadmap decisions. For growing companies, staggering commitments can also reduce the risk of being locked into assumptions that no longer fit six or twelve months later.