"It will be cheaper than the data center. You will only pay for what you use. You will say goodbye to over-provisioning."

How many times have you heard these promises? And how many times have you found yourselves looking at your cloud provider’s invoices (or those of multiple cloud providers) growing month after month without being able to understand why?

In 2025, migrating to the cloud is no longer a strategic choice, it is a prerequisite for remaining competitive. But for many companies, cloud computing has failed to deliver on its core promises. Instead of the long-awaited economic efficiency, they have found themselves dealing with:

• Unpredictable costs, with fluctuations of 30-40% from one month to the next
• Unclear invoices, where almost no one can explain why a given resource costs so much each month
• Over-provisioning that in some cases is even worse than in the data center (but at least in the data center they knew how much they were wasting)
• Conflicts between Finance and IT teams blaming one another while costs continue to rise

FinOps is not just another buzzword that companies need to take a position on, but the latest attempt to preserve the original promise of the cloud. But it only works if organizations are willing to face reality.

1. Cost management is a Finance problem

The first critical mistake companies make is assuming that cloud cost management is a financial problem. It is not. It is an architectural problem disguised as a budget problem. This is not an opinion, but a fact supported by the experience of those who have optimized hundreds of invoices from the leading cloud providers.

Why do so many companies fall into this trap? Finance looks at the invoice and sees numbers. Technical teams look at the architecture and see decisions. Cloud costs are the result of architectural choices: which instance to choose, how to structure data, when and how to scale, how to manage traffic.

The Finance team can receive and read the invoice, but it cannot modify a Lambda architecture that generates 100MB of logs per request.

Why keep asking the CFO to solve problems that only the CTO can address?

The solution, brutally simple, lies in properly assigning responsibilities and maintaining continuous communication:

• Assign cost responsibility to the Engineering team, supported by Finance
• Finance should focus on contracts, unit economics, and forecasting
• Engineering should focus on architecture and optimization

2. Cost optimization and cost management are the same thing

Stop confusing cost optimization with cost management: it is like confusing a balanced, well-planned diet with skipping a meal to feel better about yourself. Most companies focus solely on cost management when they should also combine it with deeper cost optimization efforts.

Cost management is the first step: understanding where, by whom, and how money is being spent:

• resource tagging
• monthly cost dashboards
• alerts for budget overruns

Cost optimization is the next step, going as far as deeply reviewing architectural choices:

• redesigning infrastructure and applications
• moving to serverless services
• right-sizing

True FinOps integrates both approaches, but places the emphasis on optimization. It is not about how much you spend, but how much value you get from that spending.

3. Organizational silos cost money

"When organizations don't communicate across teams, share their business context, and work toward common goals, their cloud bill shows it." (FinOps Foundation)

A typical scenario you will probably recognize:

• The CFO wants visibility and maximum cost predictability
• The CTO wants autonomy and development speed
• The product team wanted the new features ready yesterday
• The Board wants everything at zero cost, or close to it

The consequences? Each team optimizes costs and activities to achieve its own objectives, ignoring the impact on the rest of the organization. Finance cuts the budget without fully understanding the technical implications. Engineering makes architectural choices without considering their financial impact. The result is a cloud bill that tells the story of a company that does not know where it is going.

The (not so) magical formula for escaping silo hell:

1. Create a common language between Finance and Engineering teams
2. Establish shared goals that align the different business functions
3. Implement regular communication processes (not only when the bill explodes!)
4. Every architectural decision should always consider the expected and actual financial impact

4. Internal policies are full of gaps

Almost every company has detailed policies for purchasing a stapler, but allows developers to launch GPU instances with no cost controls whatsoever. This is the paradox of corporate policies in the cloud era.

The problem is that cloud IaaS and SaaS are still managed with a data center mindset and approach. In the data center, once a server was purchased, the cost was known. In the cloud, by contrast, every API call, every transferred byte, every microsecond of compute has a price.

The solution is to integrate the FinOps approach into existing policies. But how?

• Involve Procurement and Software Asset Management in cloud governance
• Create approval processes for spending on resources that exceed defined financial thresholds
• Implement automated alerts for anomalous spending
• Make every technical team accountable for the costs it generates

This is not about adding bureaucracy to innovation, but about making the financial consequences of technical decisions visible.

5. The antidote is Peer Review

How many times has a code review saved us from disastrous bugs reaching production? The same principle applies to cost-related decisions. A good peer review provides greater assurance and helps prevent costly mistakes.

Before deciding to implement:

• a new microservices architecture
• a NoSQL database
• an ML pipeline processing terabytes of data

It is always worth asking someone with a non-technical background: "What do you think this is going to cost us?".

This is not bureaucracy, it is responsible engineering.

A structured peer review process for costs:

• Prevents tunnel vision that ignores cheaper alternatives
• Shares knowledge about cost optimization best practices
• Creates accountability in architectural decisions
• Dramatically reduces billing surprises

6. IT knows everything

Technical teams are a goldmine of expertise. They often already know why that Lambda costs so much, which service is wasting bandwidth, and where over-provisioning is happening.

But when was the last time they were involved in a discussion about costs? How often were they involved in deciding whether or not to invest in a refactoring project?

How to optimize IT’s contribution in this context:

• Consult technical teams BEFORE making significant changes to the cloud budget
• Actively involve technical teams in defining cost metrics and KPIs
• Leverage the expertise of teams when evaluating alternative solutions
• Encourage technical teams to collaborate in creating cost-effective architectures

7. Cost management tools are often useless (and they cost money)

"Every byte counts, every AZ transfer costs, and somewhere, a cloud service team is probably inventing a new way to charge you for something you thought was free." (Eric Pullen — Cloud Economist & FinOps Expert)

Most cost management tools from cloud providers clearly show which resources contribute to the bill, but they do not explain the context or the value. What is really needed?

• A tool that automates analysis, not just visualization
• Dashboards that provide operational insights, not just colorful charts
• Intelligent alerts that distinguish between real anomalies and background noise
• Tools that integrate into engineering processes, not isolated tools for the Finance team
• Tools that help less technical areas of the business manage invoices and allocate costs to the correct cost centers

Of course, all of this must necessarily be accompanied by human analysis and management. A tool can enable better decisions, but human input and judgment remain irreplaceable.

So what should we do? Give up?

We have seen that FinOps is not a trend, a quarterly project, or a one-shot activity. It is a practice that must be cultivated and sustained within the organization.

The FinOps approach is the only way to truly deliver on the promises of the cloud in a world where every technical decision has immediate financial consequences.

The reality is that companies that do not choose to implement FinOps properly will continue to pay the financial consequences of every poor or reckless architectural decision.

Fortunately, there is also good news. With the right combination of:

• Clear ownership
• Structured collaboration processes
• A tool that enables insights, not just basic reporting
• Adopting a culture of financial accountability in technical decisions
• And, if necessary, a partner to support you in your decisions

It is possible to transform the cloud from an incomprehensible cost into a measurable competitive advantage.

The question to ask is not: "How can we reduce cloud costs?" but "How can we optimize the value we get from the cloud?"

The cloud is no longer just a technical matter. It is a strategic lever. And FinOps is the common language that makes it work.