Who should own cloud costs?

As the adoption of cloud technology continues to skyrocket, so does the associated spending. But cloud technology was supposed to save money. What’s happening and why?

The exponential growth of AI and generative AI initiatives are often identified as the true culprits. Although packed with potential, these advanced technologies consume extensive cloud resources, increasing costs that organizations often struggle to manage effectively. The main issues usually stem from a lack of visibility and control over these expenses.

The problems go beyond just tossing around the term “finops” at meetings. It comes down to a fundamental understanding of who owns and controls cloud costs in the organization. Trying to identify cloud cost ownership and control often becomes a confusing free-for-all.

Why are we having this discussion?

A recent CloudZero survey revealed that a staggering 58% of respondents believe their cloud costs are excessively high. (I read a pretty good follow-up article by the senior VP of engineering at CloudZero.) Of even more concern is the fact that only 42% could estimate how much of their cloud spend goes to various parts of their business. The result? Organizations are financially compromised and suffer a direct hit to productivity due to this cloud cost opacity.

The study also highlights that engineering is particularly affected. Two-thirds of engineers reported disruptions caused by poor visibility into cloud costs. Apparently, there is a disconnect between engineering efforts and business outcomes. It’s a scenario no business wants: misaligned priorities leading to inefficiencies and wasted resources.

CloudZero agrees that pushing the engineering team to own and control the costs results in better outcomes. The same CloudZero survey found that 81% of respondents had cloud costs “about where they should be” when the engineering team held some level of responsibility for these expenses. Indeed, they consider this an insight into the transformative potential of proper ownership structures.

Why does giving engineering control over cloud costs make such a difference? For one, engineers are typically closer to the actual usage and deployment of cloud resources. When they build something to run on the cloud, they are more aware of how applications and data storage systems use cloud resources. Engineers can quickly identify and rectify inefficiencies, ensuring that cloud resources are used cost-effectively. Moreover, engineers with skin in the game are more likely to align their projects with broader business goals, translating technical decisions into tangible business outcomes.

What do the engineers think?

On the other hand, I’m often alerted to excessive cloud costs incurred by, you guessed it, the engineering teams. Engineers definitely have the potential to be the good guys of cloud computing who use these resources more cost-effectively, but most engineers are not born that way. Implementing this shift requires more than just a mandate from above; it requires interdisciplinary alignment among engineering, finance, and infrastructure teams. These groups must share a common understanding and a unified strategy around what constitutes “cloud efficiency.”

The first step is to create a single source of truth for all cloud cost data (PaaS and SaaS), to ensure consistency and transparency across departments. This means employing finops or finops-like systems that can track and report on cloud costs by user, department, use case, profit center, etc.

CloudZero’s argument that engineers should own and control costs would require them to own finops as well. However, if I were to call all my clients’ engineers into a room and inform them that they now own all costs, they would find that announcement hilarious. It doesn’t matter if most would manage their costs efficiently; total cost ownership by engineers is a bridge too far.

Engineers build things. I want them to build things—and architects, too. They may have better insights into costs, but my preference would be for them to work closer with the finops team to assist them in understanding how the cloud resources are being used. This would require some training, and I would have to explain to them why this is a good thing. In many instances, I suspect we would end up in an argument as the engineers question why these additional cost-monitoring responsibilities have been thrust upon them. These are conversations I would rather not have.

We have to live in reality. I agree with the report and the article that engineers would be great teams to monitor and own cloud costs. Indeed, they are closest to the point where the costs are better understood and controllable. However, I would also love the Easter Bunny to bring me candy once a week. Enough said.