Like many things in life, when you’re new to the cloud you don’t know what you don’t know. Given that migrating workloads to the public cloud is often a key component of a business transformation initiative, you want to avoid a long, expensive learning curve—especially since accelerating time-to-value is often a major impetus for the move. We recently wrote about widespread overconfidence in existing cloud tools and capabilities that we found in our recent survey of hybrid cloud decision makers. As they say in late-night informercials, but wait—there’s more.
When we uncovered the incongruity between confidence levels and actual abilities and results, we took a closer look at whether the number of workloads respondents had already migrated to the cloud, a proxy for experience, made a difference. Spoiler alert: it does.
The hybrid cloud has a learning curve
Given the complexity of cloud migration and deployment, it should come as no surprise that organizations will likely go through a learning curve. But what exactly does that mean? Are there some areas that have a steeper curve than others? As we analyzed the survey results through the lens of experience, we found evidence of learning curves in three key areas: planning and forecasting, understanding and tracking actual spend, and adapting and optimizing over time.
Planning and forecasting
There’s a learning curve when it comes to forecasting usage patterns, but it’s not too severe. Respondents with less than one-quarter of their workloads in the cloud, however, are far less likely than their more experienced colleagues to be able to predict long-term spending needs—68% vs. 82-83%. This means that while organizations new to the cloud seem to have a decent handle on what their workloads will be doing, they’re less likely to understand the cost ramifications of those forecasts, which could result in higher-than-expected spend.
Understanding and tracking actual spend
One way to understand your cloud spend is to analyze your cloud bill, but this can be easier said than done. While the majority of respondents say they can do this, only those with more than three-quarters of their workloads in the cloud can do so nearly universally (94%). For the others, 16-17% don’t have any capability. There’s a lot more detail around spend that organizations likely want to understand beyond the information available in the monthly cloud invoice. Only 63% of respondents with the fewest number of workloads currently in the cloud have this ability, compared to 77% for organizations with 26-50% of their workloads in the cloud, and 83-85% for those who are more than half cloud-deployed. Those who have not yet migrated to the cloud may want to reassess their expectations as they may be overestimating their ability to get the cost details they need.
Adapting and optimizing over time
Because the cloud is a dynamic environment, you need to understand cost-impacting changes and then be able to adjust accordingly. Here again, the more workloads a company has in the cloud, the better able it is to respond to shifting conditions and requirements—mostly. Only 63% of respondents (those with up to one-quarter of cloud-deployed workloads) said they can adjust spending as needs change, compared with 90% of those that have half to three-quarters of their workloads in the cloud. Interestingly, this number dips to 83% among those who have deployed more than 75% of their workloads in the cloud. We can only speculate as to why this may be the case. Perhaps these organizations employ more reserved instances, or maybe the sheer volume of workloads they’re running in the cloud creates a management complexity that impinges on agility. Here, too, when it comes to organizations who are still in the cloud planning stages, we see a much higher level of (over)confidence than those enterprises who have started executing. Finally, to react quickly, you need real-time cost alerting, but only 55% of those in the 1-25% cohort have this capability, compared with 77% in the 26-50% group, 83% in the 51-75% group, and a whopping 96% among those with more than 75% of their workloads in the cloud.
Lessons can be expensive…
All of this matters because if you can’t effectively forecast, track, and adjust, you will end up needlessly spending a great deal of money. In fact, 82% of survey participants told us that they had incurred unnecessary cloud costs. There were a variety of reasons why this happened, including workloads bursting above agreed capacity (41%), overprovisioning and overbuying (35% and 34% respectively), unattached storage blocks (34%), and even poor job scheduling (22%). Compounding the problem is the fact that most of these organizations squandered money on multiple fronts, with 29% of respondents citing two factors contributing to unnecessary costs, and 32% experiencing three or more.
…but it doesn’t have to be this way
Migrating workloads to the public cloud is a big step—one with its own best practices and challenges—but it’s only the first step. What happens on day two and beyond requires vigilance and proactivity to ensure those workloads operate at optimal levels at the lowest possible cost. If you’re figuring out how to do this via trial and error, building knowledge and expertise over time, chances are high that you’ll end up spending a lot more money than necessary, and may even suffer performance problems and other issues in the process. Fortunately, there’s an alternative. With Virtana Platform, you gain precision observability via the combination of AIOps, machine learning, and data-driven analytics for efficient migration and ongoing performance, capacity, and cost optimization—not to mention the experience of a partner who’s been doing this for many years. As a result, you’ll know before you go, and you’ll get and stay rightsized. Request a free trial today!
Get the survey report: State of Hybrid Cloud and FinOps survey
Request a trial: Virtana Platform
Ricardo Negrete
