Why Most Cloud Business Cases Get It Wrong

Part 1: The Pitfalls That Undermine Cloud ROI

By Lou Perugini | CloudTechCFO.com

The CFO’s Cloud Paradox

Cloud is often sold as a no-brainer. Lower costs, more agility, faster innovation. What’s not to

like?

But for many enterprises, cloud investments fail to deliver the returns they were promised. The

forecasts looked good on paper. The migration went fine. The invoices, however, tell a different

story.

As someone who’s reviewed hundreds of cloud business cases and worked with finance and

technology leaders across industries, I’ve seen a pattern:

Most cloud ROI models fall apart not because the technology failed, but because the

assumptions were wrong.

This article unpacks the most common pitfalls I see in cloud business cases. In Part 2, I’ll show

how CFOs can fix the modeling logic, align spend to outcomes, and turn cloud into a reliable

driver of business value.

1. The CAPEX Mindset Still Lingers

Most enterprises controlled on-premises IT spend through CAPEX approvals and purchase

orders. That process served as a built-in governance mechanism. Spending was approved up

front, and changes were rare.

Cloud breaks that model.

With cloud, there are no POs, no depreciation schedules, and no hardware refresh cycles to

manage. Anyone with a company credit card and minimal permissions can spin up infrastructure

in minutes.

The old gates are gone—and with them, the controls.

Without rethinking governance, many companies unintentionally recreate the same “free for all”

dynamic they used to attribute to shadow IT.

2. Pent-Up Demand Comes Flooding Out

Most IT organizations are unaware of how much unmet demand for technology exists inside the

business. There’s no line item for it, no budget request, and no champion—it’s just a backlog of

needs that haven’t been addressed due to limited capacity.

The moment you move to cloud and remove provisioning bottlenecks, that demand explodes.

You’re not just funding a migration. You’re unleashing a wave of experimentation, self-service,

and growth initiatives that never had a shot before.

This demand surge is healthy, but if you don’t anticipate it, your cost forecast will implode.

3. Lift-and-Shift Is a Waste Multiplier

One of the most common mistakes in cloud migrations is recreating the on-prem environment in

the cloud, without questioning whether it makes sense.

If a server on-prem had 16 cores and 128GB of RAM, many customers spin up a 16-core,

128GB VM in the cloud—because it feels like the “safe” move. But most on-prem workloads are

massively underutilized. During my time at AWS, the average on-prem server we analyzed ran at

just 39% CPU and 42% memory utilization. Many were in the single digits.

Lift-and-shift often turns underutilization into waste at scale.

Cloud requires a new playbook: one that prioritizes auto-scaling, performance-based rightsizing,

and horizontal architecture choices that don’t simply mirror legacy infrastructure.

4. Legacy Costs Don’t Go Away on Their Own

Here’s a big one that often gets left out of the case altogether:

You’re not saving money if you’re still paying for both environments.

It’s tempting to build a cloud ROI model that focuses on the new cost profile, but unless there’s a

detailed and aggressive plan to decommission legacy infrastructure, real savings won’t

materialize.

Stranded assets, unused data centers, duplicate licenses, and residual contracts quietly eat away

at your margins. Even worse, the headcount and operating overhead to manage both

environments often stays in place longer than expected.

Every month of overlap erodes the business case.

5. Forecasting Behavior Like It’s Hardware

Another trap: trying to model cloud usage like it’s fixed infrastructure.

On-prem cost planning is relatively straightforward—buy the server, depreciate it over five

years, and maybe renew the software license in year three. Cloud is different. It’s variable, real-

time, and human behavior-driven.

Without strong usage forecasting discipline, cloud costs can grow quietly and unpredictably,

especially when innovation teams are building quickly and tearing down slowly.

Cloud economics are not just financial. They are behavioral.

A sound business case needs to account for ramp-up periods, scaling profiles, seasonality,

and—importantly—the learning curve of the teams consuming cloud.

6. No One Owns Optimization

The initial business case usually assumes that workloads will be optimized. But unless that

optimization is assigned, incentivized, and governed, it rarely happens.

Many organizations don’t have a FinOps practice in place when cloud adoption begins. They rely

on a single cost analyst, a few dashboards, and hopeful engineering habits.

In reality, optimization must be a distributed team sport. That means:

•  Engineers get access to cost data

•  Product owners understand tradeoffs

•  Finance partners with tech on budget accountability

•  Leadership sets goals for unit economics, not just total spend

Without this structure in place, cost sprawl sets in, and savings stay theoretical.

7. The ROI Model Ignores the Real Win: Enablement

Perhaps the biggest blind spot in cloud business cases is focusing too much on cost savings and

not enough on enablement.

Cloud is not just about reducing IT expense. It’s about enabling faster product delivery, running

more experiments, automating more finance functions, and unlocking new revenue streams.

These benefits are harder to quantify, but they are often where the true ROI lives.

When CFOs pressure-test cloud cases, they need to ask:

What business outcomes will this enable that we couldn’t do before? How soon will we

know if they’re working?

If your case only reflects infrastructure savings and doesn’t capture acceleration, agility, or value

unlocks, it’s incomplete.

Coming Up in Part 2: The Fix

These pitfalls are common, but they’re not inevitable.

In Part 2, I’ll walk through the CFO playbook for building cloud business cases that actually

work. You’ll learn how to:

•  Shift from total cost to marginal value

•  Fund MVPs, not monoliths

•  Integrate legacy wind-down into your ROI

•  Build staged investment and governance into your model

•  Use FinOps principles to create budget accountability

•  Translate enablement into measurable outcomes

Want it sent straight to your inbox?

Contact me BELOW or follow me on LinkedIn to get notified when it drops.

LET'S TALK