Was moving to cloud a mistake?

According to a new study from Wanclouds, 81% of IT leaders say their executives and boards of directors have directed them to reduce or take on no additional cloud spending. It’s little surprise as cloud costs skyrocket while the economy remains in flux. After multiple years of unimpeded cloud growth, the findings suggest that enterprises’ unbridled cloud spending may soon temper. Is this true?

I’m always skeptical of any study that says something extreme is happening. While aspects of these studies are often true, the overall technology market shifts slower than most people understand. Enterprises will take their own sweet time to slow down or speed up anytime technology is involved.

Also, you need to look for the underlying motivations and conflicts in all these studies. For instance, you won’t see a security company sponsor a study that finds everything is as secure as possible and nothing can be improved. That’s not the case in the Wanclouds study, but please be on the lookout for bias, especially when studies show unexpected results with large swings in one direction or the other. 

In the case of the Wanclouds study, there is truth because two things are occurring in the market right now. 

First, enterprises are hitting a “complexity wall.” They have so many services onboarded that they have no way to operationalize them within existing budgets and resources. According to the study, “Multicloud usage is becoming increasingly unwieldy, and costs are difficult to manage across hybrid environments.” I’ve talked about those problems before, many times, so I won’t belabor the point. 

Second, as I recently covered, ongoing cloud costs are shocking most enterprises. With barely 20% to 30% of enterprise workloads on the public cloud, the bills are much higher than expected. Part of this is a lack of planning which results in underoptimized cloud solutions. Also, there is a lack of accountability and spending discipline, meaning that finops is nowhere to be found. 

Most cloud ROI problems are self-inflicted. However, the industry bears some culpability for overselling and overstating cloud cost savings. In general, when I’m called in to help fix a cloud implementation gone wrong, I find two root causes: First, little thought went into the planning that needed to occur before the first purchase decision was made. Second, trying to lift and shift your way to success rarely works out.

It’s tempting to wag fingers at enterprises that got into cloud trouble, but that’s not productive. It’s better to determine where your enterprise is on its trip to the cloud and then figure out how to incrementally improve in both the short and long term.

You may have to go slower to go faster. Some missteps need to be fixed, such as massive lift-and-shift projects that moved poorly designed and built software. Migrating software that didn’t work well in the data center will not magically solve bad design once it’s in the cloud. No cost improvements will be found at the end of that journey. The result, in many instances, is more complex deployments that cost 30% to 40% more to operate on the cloud, all in. These problem programs need to be refactored before migration, fixed on the cloud after migration, or rewritten to take advantage of cloud-native features (which is where the largest hard and soft cost savings reside). Or research might reveal that the best operational and most cost-effective solution for a particular application is to remain in the data center.

Focus more on planning and deployment. Use optimized architectures instead of what seems to work or what someone else hypes. Containers, for instance, are solid solutions for existing and new applications, but they should be evaluated along with all other alternatives, such as using traditional non-container dev. Operational complexity should be factored into the solutions. The goal is to move to the most optimized solution for simplicity and cost. Sometimes that involves something different from what we envisioned. 

Adjust expectations. Cloud computing is not the savior of poorly run IT. It’s just another enabling technology that works well—if the right amount of planning occurs before resources are committed. In the early hype-filled days ten years ago, this reality was often shouted down. The situation we’re in now is not some “I told you so” moment that vindicates those of us who were shouted down; it’s a clear and present wake-up call that we need to learn from poor assumptions made in the past and make better plans for the future.

Yes, there will be some uncomfortable conversations with your executive team and board of directors about some of the issues arising. The key to success moving forward is to admit that things need to change and to have the plans in hand that show you are willing to do the heavy lifting to make changes. 

We’ve been through these cycles before. This time it will take more time and money than other past improvements and fixes. Cloud complexity is a problem many didn’t see coming. It’s here to stay. New tools and configurations come onto the cloud scene almost weekly to help deal with complexity, but underlying problems still need to be addressed. We hurried up to get to the cloud and the lack of planning is starting to show. If you hurry up to apply an endless parade of fixes, that just kicks the can down the road. It’s time to take a step back, identify the problems, do your research, and create a plan to fix the problems. I can see no other choice.