Visibility into cloud costs is a challenge for most organizations, with engineers and finance teams struggling to correlate their investment in cloud technology with the value it brings to the business.
This was among the findings of a CloudZero survey, which asked more than 1,000 engineering and finance professionals to weigh in on the various dimensions of their cloud cost strategies.
Overall, the study found organizations with engineering cultures of cloud cost ownership have:
- improved visibility and cost attribution
- higher confidence in financial reporting
- better adherence to budgets
Nearly two-thirds (65%) of respondents said they feel cloud cost is a shared responsibility between engineering and finance.
Related: 5 Myths About Cloud Costs
Among the other key findings from the study is the fact that just 13% of organizations say they have allocated more than 75% of their cloud costs.
Cloud Costs: A C-Suite Priority
There is broad consensus that the cost of cloud technology needs to be a top issue for the executive team, with nearly three-quarters (73%) of respondents agreeing cloud cost is a C-suite or board-level issue.
Just under half (49%) of respondents said cloud costs are higher than they should be, with 11% agreeing that cloud costs are "way too high."
Erik Peterson, CTO and founder of CloudZero, said to improve cloud cost visibility, the first thing engineering and finance teams need to do is work together.
"Engineering holds the keys to cost visibility and control, while finance benefits the most from accurate reporting and budgeting," he said. "It needs to be something that's proactively discussed and collaborated on from the beginning. Finance shouldn't be running over to engineering at the end of the month, shocked about the state of the bill."
With that in mind, however, Peterson noted that finance doesn't care about how much money was spent on EC2 or Kubernetes this month; they need to understand how the costs map back to the business — this is the secret to getting engineering and finance to work together.
"Use terms that are tied to the unit economics of your business, like cost per customer, transaction, product, or feature," he said. "Now you have both teams speaking the same language."
Needed: A Culture of Cost Awareness
The second key point is that engineering teams need to foster a culture of cost awareness, Peterson said.
"Over the past decade, engineering teams have embraced disciplines like security and QA, incorporating them earlier in the software development lifecycle, and cost should be no different," he said. "Considering cost earlier and throughout the SDLC leads to better outcomes for everyone — and the survey data proves that."
Related: How Green Is the Cloud?
Davis McCarthy, principal security researcher at Valtix, a provider of cloud-native network security services, pointed out that because the cloud provides flexible compute resources, organizations are no longer bound by their on-premises hardware.
"Lifting this barrier allows for innovation to spike, sometimes resulting in unexpected cloud costs," he said. "However, platforms like AWS have a robust tagging system for their resources."
Tagging is part of cloud cost allocation, which involves identifying, aggregating, and allocating cloud spend among specific use cases within a company.
Creating a tagging framework that works for engineering and finance will make identifying resources and their costs easy, McCarthy said.
"Consider the scenario where customer growth is responsible for an increase in cloud costs. Finance now has the information that cloud costs have a linear relationship to user activity, and engineering shouldn't lose out on their budget to innovate because of it," he noted.
McCarthy said the benefits of tagging also include improved asset management and security policy, a cross-team win for those using the cloud.
Knowing what it costs to support a feature can help organizations set correct pricing for strong margins and profitability.
In addition, cloud cost allocation planning means being able to identify which customers are the least and most expensive to host and help determine if the organization should revise the pricing for the expensive ones (or specific customer segments) to improve gross margins.
"As new tools become available in the cloud, more unknowns will exist," McCarthy said. "Emerging threats resulting in the abuse of those tools, scaling and efficiency problems, as well as growing user inexperience may result in unexpected costs."
Peterson pointed out that modern companies are building architecture that is highly scalable and ephemeral.
"Cloud-native designs like multi-tenancy or Kubernetes abstract software from the infrastructure it's running on — which in turn misaligns how cloud providers bill from the way organizations think about their business," he said. "Companies need to find ways to understand the cost of what's running on their cloud infrastructure, not just the infrastructure itself."