Many organizations move to the public cloud expecting lower total cost of ownership, but the savings often don’t show up because of how they migrate. The most common issue is a **“lift and shift” or direct-match migration**—moving existing servers, storage, and applications to the cloud *as they are*, without rethinking capacity needs. Key problems highlighted in the data:

• According to an analysis of **more than 60,000 workloads** by Bain & Company and AWS, a poorly planned direct-match migration can be **10%–15% more expensive** than keeping workloads on-premise.
• Using AWS Migration Evaluator, they found that **84% of on-premise workloads are overprovisioned**—meaning companies are paying for excess compute and storage they don’t actually use.
• When this excess capacity is moved unchanged to the cloud, companies simply **transfer their inefficiencies** to a new environment.
• In a survey of **more than 350 IT decision makers**, the top disappointment with cloud deployments was that **cost of ownership had not declined**.

In short, cloud migration doesn’t automatically lower costs. Without upfront analysis and rightsizing, organizations often pay more for the same (or unused) capacity in a different place.

In cloud migration, **rightsizing** means rethinking and adjusting your workloads so you only provision the compute, storage, and licenses you actually need—rather than copying your on-premise setup directly into the cloud. This involves:

• Identifying and eliminating **zombie servers** (systems no longer needed but still running).
• Shutting down or winding down **unnecessary activities** and duplicate workloads.
• Analyzing **usage patterns** (average and peak demand) to size instances more accurately.
• Using **burstable instances** or a reserve of excess capacity to handle demand spikes instead of overprovisioning all the time.
• Rationalizing **core processor counts** to optimize software licensing fees tied to cores.
• Reevaluating **storage patterns** (average vs. peak usage, and proximity to compute) to downsize storage tiers appropriately.

The impact on cost can be significant:

• Bain & Company’s experience shows that companies that rightsize their workloads can **cut migration-related costs by about 30%–60%**.
• Organizations with existing **Microsoft SQL licenses** that can be migrated often see the highest savings, but even those without transferable licenses can still generate substantial reductions through rightsizing alone.
• In one AWS Migration Evaluator example, a direct-match migration of **105,000 server instances and 20 TB of storage** would have **increased costs by 22%**, while rightsizing the same workloads would have **reduced costs by 36%**.
• Morningstar, a financial research firm, found that after rightsizing, **the majority of its instances would cost about 30% less** in the cloud than on-premise.

Rightsizing is essentially using migration as a chance to **reimagine your infrastructure footprint**, rather than replicating it. That’s where most of the savings come from.

Not all workloads deliver the same economic benefit when moved to the public cloud, even when you rightsize. The data suggests some clear priorities. 1. Focus first on nonproduction workloads

• Nonproduction environments (development, testing, staging, etc.) usually have **more volatile and variable usage**.
• This volatility makes them better suited to the **on-demand flexibility** of the public cloud.
• For these workloads, migration can also help reduce **software licensing fees tied to core processor counts**, because you can more tightly control and rightsize the cores you use.

2. Target nonvirtualized workloads before virtualized ones

• Nonvirtualized workloads typically gain more from cloud migration because they don’t already benefit from the efficiency and flexibility that virtualization provides.
• Virtualized environments have already captured some of the same benefits the cloud offers, so the incremental savings from moving them can be smaller.

3. Use migration to clean up and consolidate

• Many companies discover **fragmented and duplicated workloads** scattered across machines and data centers.
• They often uncover **incomplete or outdated records** and **zombie servers** that can be decommissioned instead of migrated.
• By consolidating and rightsizing these workloads before or during migration, companies can materially lower their cloud footprint.

4. Think beyond cost: hybrid and vendor opportunities

• Rightsizing also helps organizations manage a **hybrid environment** more effectively, balancing public cloud with on-premise systems.
• Technology vendors can play a role by helping customers **scale, automate, and operate** across public, private, and hybrid models, as seen in moves like **IBM’s acquisition of Red Hat**.

In practice, a rightsized strategy usually starts with **nonproduction and nonvirtualized workloads**, uses migration as a chance to clean up and consolidate, and then extends to more critical production workloads once the economic model is proven.