I doubt whether anyone hasn’t read technical media and financial articles that talk about the slowing growth of cloud computing. I also doubt anyone has missed the stories of cost overruns and companies who have saved big bucks not by adopting the cloud but by moving back to, or in some cases even creating, data center hosting. I’ve blogged in the past about the reasons why all this is happening, but there’s an open question on just how much of that users are actually aware of. Are enterprises overall seeing cloud issues, and are they doing something about it? If so, what does this mean in terms of cloud adoption.
Cloud adoption, according to almost 90% of enterprises I hear from, is driven by application modernization, which in turn is almost always linked to making applications accessible via the Internet as web sites, apps on smartphones or tablets, or both. This process has been going on for a decade but accelerated at two specific points, one seven years ago to respond to exploding online sales interest and one during COVID to respond to lockdowns and work-from-home (WFH). Most (54%) of enterprises say that “the majority” of their appmod activities are complete, which is why cloud growth rates have slowed.
The second truth here is that virtually every enterprise that uses the cloud today admits that their cloud projects were subject to major time constraints when launched and as they evolved, and well over three-quarters of enterprises believed that this resulted in less-than-optimum application designs and cloud feature and even cloud provider selection. Of 197 enterprises who offered cloud views in the last two months, only nine said they believed they were fully satisfied with their cloud designs and decisions. In contrast, eleven said they believed they should not have used the cloud at all.
The thing that’s driving what we might term “cloud reconsideration” is the orderly updating of the applications that were supported via the cloud. Nearly all enterprises who adopted cloud computing through appmod said that they made “at least annual” tweaks to the applications to respond to things they learned or to changes in their business. When these tweaks were considered, development and operations reviews often spotted issues with application performance, efficiency, or both. But enterprises also say that application and cloud service changes in response to these issues were triggered by the cost-management issues that their businesses faced in 2023. That’s why so many of the stories on cloud growth and spending emerged only this year.
Cloud growth, even net cloud spending growth, hasn’t gone negative, it’s just slowed. That’s partly because most enterprises still have appmod work to do, and even that 54% who say a majority of their appmod is done still have, on the average, about a third remaining. The rest have more than half, and so there’s still a fertile field to cultivate here. However, the big adopters of the cloud are nearly all in that 54% who say the majority of their work is done. That suggests that cloud growth rates may recover a bit as the negative impact of cloud reconsideration fades next year, but that they probably won’t resume the upward trajectory they had prior to 2023.
Well over two-thirds of companies say that their cloud spending even in 2023 will be higher than that of the prior year, but almost the same number say that their growth rate in cloud spending will decline in 2023 and that they expect it to decline again in 2024. A bit under two-thirds think cloud spending in 2025 will grow more slowly than in 2024, which shows that there’s a general sense that they’ve spent too much on the cloud.
Why? A variety of reasons are offered, with most companies offering more than one. The top ones are cloud pricing is too complex (78%), bad application design choices (69%), usage was higher than predicted (61%), operations and deployment need optimization (55%) and changes to application raised costs (51%). Only 35% of companies said their cloud decision-making process overall was flawed, even though my own view is that this was the root cause of most of the other issues. Nobody wants to say they messed up the whole project, though most do to a degree. Blaming others is OK; 40% said that cloud coverage in the tech media misled them, 36% said that the cloud provider misled them, and 17% said that professional services advice from a third party was a reason for a spending overrun. One common point was that 82% said that they didn’t really have enough cloud expertise when their initial decisions were made. Only 14% think that’s still a problem.
Some other data points from the last two months offer some useful correlations. Of the technical people who I emailed with, only 17% said they believed their company would “move to the cloud”. A little over half said that there were some applications they believed would be moved to the cloud, but only a quarter said that any “core” applications would be moved. The majority of the candidates for movement were applications relating to payroll, personnel, sales management, and CRM.
Multi-cloud usage was reported by 32% of companies, but almost 70% of those said one of the two clouds was a SaaS application set, and a quarter considered Microsoft 365 to be a cloud/SaaS application. That means that among these enterprises only about one in five was “multi-cloud” in the sense that we usually expect the term to mean. While almost half had geography-related reasons, the largest driver was application-specific cloud features that all providers didn’t offer. Remember that appmod is the big driver of cloud usage and appmod tends to look at one application at a time, so any mandate to use a single cloud for multiple applications either has to be set as a policy in advance (companies generally say they have no such policy) or imposed by CIO or CFO in response to cost or other issues (only a few companies say this has happened).
I continue to be surprised and confused by the gulf between the way that cloud adoption and usage is described and the way users say it’s actually evolving. Perhaps I’m biased here, but I believe that the reason for this is that coverage of the cloud is driven by a combination of vendor/provider initiatives and search engine optimization. The former obviously tends to favor vendor goals, and the latter is influenced by the fact that the majority of searches on cloud-related terms and concepts is done by people looking to be educated in a professional sense rather than people actually preparing for cloud adoption or refining their cloud plans. Whatever the cause, we do seem to be slowly converging on a set of cloud axioms that represent what’s really happening, which is a good thing.
