The rise of AWS has been rapid across all industries. In 2019, AWS achieved sales of over $35bn – 36% up on the previous year. This growth shows no sign of slowing down, with some analysts predicting AWS revenue will surpass $100bn in 2022. AWS’s success has been driven in part by the growing demand from organizations for elastic and scalable technology cost models.

These have sought to break away from traditional procurement approaches, with their periodic capital expenditure and requirement for forward-planning that demands high degrees of perception – and often luck – to strike the right balance between having room for growth and minimizing unused capacity. When they first appeared on the scene, cloud providers such as AWS, helped address this need. They offered on-demand infrastructure, which promised to enable organizations large and small to innovate faster, respond more effectively to opportunities and threats, and reduce costs.

The time of the IT opex model had arrived. There was great fanfare about the savings organizations might be able to achieve. And indeed, many did cut their IT spend significantly, particularly where they were able to fully vacate legacy data centers. As with so many things, however, the devil is in the detail.

It starts out fine

One of the strengths of cloud services is that developers, database administrators and other IT professionals can try things out by quickly spinning up resources and billing them to a company credit card. In the early days, when total use of AWS in an organization is relatively small and noncomplex, keeping a handle on the technical and commercial make-up of the estate, and the workloads running on it, is manageable. Equally, any financial losses resulting from non-optimized resources (such as over-sized instances, or environments left running when they’re no longer needed), while undesirable, do not represent a significant part of an organization’s overall IT costs. In situations where cloud usage remains relatively low, the cost of eradicating this kind of inefficiency can be higher than the savings that would result.

The challenges of scaling

As these early AWS workloads start to deliver success, the natural progression is for more – and more complex – workloads to follow. This is a gradual process, and typically continues on the same path as the early cloud adoption, with individuals and departments provisioning and managing their own AWS resources. In this way, the AWS estate grows, becomes more complex, and fragments. And this is where the challenges begin to arise. Firstly, keeping track of what the organization is running, where, and what for, becomes an increasing headache. And secondly, the losses from inefficiencies become more significant.

In the Flexera 2020 State of the Cloud Report1 , respondents estimated that their organizations waste an incredible 30% of their cloud spend. For SMBs, 56% of whom surveyed by Flexera said they spend up to $600K per year on cloud, that level of wastage wipes as much as $180K per year off profits. At the other end of the scale, 20% of the larger enterprises surveyed spend more than $12M per year on cloud – wastage of 30% here will be wiping an eye-watering $3.6M off their annual profits. The report also found that the two challenges we’ve highlighted above are on many IT leaders’ minds.

Respondents were asked what their top cloud challenges were. In second place was the management of cloud spend, identified by 79% of respondents, while in third place was governance, highlighted by 77%. As one would expect, these figures were higher in enterprises than they were in SMBs.

Addressing the growing challenge of governance and spend management

So how do IT leaders ensure their AWS estate remains lean and aligned to business goals, so that it continues to deliver the promised cost savings, even as cloud usage expands?

The solution is threefold. Firstly, it demands a consolidated view of the organization’s AWS spend. Achieving this requires the implementation of suitable centralized cloud spending governance, which is the second pillar. And thirdly, it necessitates ongoing optimization of the AWS estate, looking at where commercial and technical improvements can be applied, to minimize wastage and free up precious resources to be used where they’re most needed.

The third pillar is particularly important, and implementing it successfully requires a blend of expertise, to address workload architecture and platform design, as well as the many licensing challenges that can arise when running workloads in the cloud. We dive into some of these opportunities for optimization in greater detail in the second part of this series.