A survey of 315 IT professionals in the U.S. published today finds more than three quarters of respondents (78%) estimate that anywhere from 21 to 50% of their organization’s spending on cloud services is wasted.
Conducted by Stacklet, a provider of a platform for governing cloud computing environments, the survey also finds 51% of respondents estimate more than 40% of their cloud spend is wasted.
The primary reasons cited for lack of control over cloud resources include dependence on inefficient manual processes (52%), inability to control cloud usage (51%), lack of best practices (47%), limited visibility (45%) and lack of timely action by engineering teams (40%).
Stacklet CEO Travis Stanfield said much of the wasted spending can be traced back to a tendency to over provision cloud resources. Application developers that typically provision cloud resources themselves tend to request as much infrastructure resources as possible as part of a desire to ensure applications are always available. Much of that infrastructure, however, winds up sitting idle.
Unfortunately, far too many organizations lack visibility into how cloud resources are actually being consumed, noted Stanfield.
The most common sources of inefficiency cited are excessive log/data retention (50%), incorrect resource sizing/type (49%), costly artificial intelligence (AI) experiments (48%), outdated instance types (48%), misconfigured networking (48%) and application development systems running during off-hours (47%).
A full 88% of respondents encounter these issues several times yearly, with 50% experiencing them monthly or weekly, and 62% reporting mistakes costing their organization more than $25,000 per month.
In general, many IT and business leaders feel they need to make a false choice between innovation and efficiency when they can have both, said Stanfield.
On the plus side, more attention is being paid to implementing best practices during the last two years as economic growth has stagnated around the world, said Stanfield. However, the rise of AI is driving additional wastage of cloud resources, he noted. A full 82% of respondents agree AI is fueling increased cloud complexity and spending.
There may come a day when AI, of course, also makes it simpler to govern cloud spending. In the meantime, however, it’s clear that many organizations are still finding it challenging to manage these environments, which often result in unpleasant surprises when the monthly bill for cloud services comes due.
Ultimately, organizations need to be able to readily access cloud cost metrics at the time when these resources are provisioned. Most application developers once they are made aware of cost issues will make better choices. However, mistakes will always be made, so IT operations teams will always need governance tools that enable them to enforce cost control policies, noted Stanfield.
The biggest challenge might not be acquiring and deploying the tools and platforms needed to enforce those policies. After allowing application developers to programmatically provision resources on their own for the past decade, many organizations will need to change an internal culture that has not always emphasized cost control, said Stanfield.
One way or another, however, the one thing that is certain is the finance teams that pay the cloud bills when they come due are starting to ask a lot more questions about how cloud resources are specifically being consumed.
