Cloud Economics Needs a Re-education
- By Winston Thomas
- March 14, 2023
It’s every procurement and senior IT executive’s nightmare: uncontrollable cloud costs.
After gorging on SaaS and embracing the flexibility of the infrastructure, companies are now facing sticker shocks. It looks like moving to the cloud may not be as simple as it ought to be.
In fact, cloud economics is such a worry that respondents to the 2023 State of the Cloud Report by Flexera prioritized it over security for the first time in a decade.
Now companies are thinking of a U-turn. Many are considering moving back on-premises. Others are just cutting down cloud spend to stop the bleeding.
What we need is a different approach to cloud economics. But before we take it, we must first ground ourselves with some sobering facts and eliminate misconceptions.
Time to stop using the simplicity argument
The old argument was that cloud simplifies your infrastructure needs by shifting your CAPEX to OPEX, offering you room for peaks and experimentation. If only life were this simple.
This argument was designed to persuade those still stuck in an on-premises world. That’s not what we face today, where hybrid and multicloud architectures are more common. It also means the old economics model used to justify cloud migration and simplify infrastructure build no longer applies once you start migrating.
In McKinsey’s “Cloud economics and the six most damaging mistakes to avoid”, the authors claimed that historical trends no longer apply once you’ve started to move into the cloud.
“While companies make allowances based on cloud’s prevailing operating-expenditure model, old habits are hard to break, and forecasting typically still relies heavily on the capital-expenditure model. This often results in a greater than 20% discrepancy between forecast and actual spend, leading to poor allocation decisions and arduous rebudgeting,” says the article.
Part of the discrepancy is also because cloud adoption changes user behaviors, which does not really get into IT procurement cloud price calculations.
According to McKinsey, when you remove the need to actually buy and deploy hardware, a rental mentality sets in. You will see different departments asking for more. And with the cloud’s accessibility, they might even bypass IT and procurement, resulting in sticker shock.
Recalculate the cloud elasticity benefits
The ability to spin up virtual machines as needed is one of the best features of the cloud. It eliminates a significant reason why departments or companies are wary about experimenting with creating new services. It also helps marketing and sales departments to drive campaigns that drive peak activity without having a long IT procurement cycle.
But this so-called elasticity is not for all. Other parts of a business, from finance to HR, do not need this elasticity. Yet, give this option to these departments, and they will ask for it.
So, cloud economics and procurement need to get better at differentiating workloads, said McKinsey. Their article noted that “companies need to examine their workloads individually to assess whether their elasticity patterns would lead to savings on the cloud.”
Don’t forget the architecture variable
Companies tend to approach clouds from a pure economics standpoint. That’s all well and good when moving from an on-premises starting point.
But architecture can be a significant sore point once you've adopted the cloud. That’s because a successful adoption and deployment of the cloud will see many departments drive up network utilization rates and use up more IT resources.
If the architecture cannot keep up, this utilization rate can clog up pipelines and reduce user experience. This can drive up costs and unbalance the initial economic calculation.
According to Cap Gemini’s “Impact of Planning and Architecture on Cloud Economics” paper, you need to start with a high-availability design from the onset.
Old ideas of batch processing do not sit well in a cloud environment. Neither does the old one-database-for-all-apps approach. For the latter, Cap Gemini authors suggest exploring polyglot persistence.
If you’re in a greenfield scenario where you are using the cloud to design a brand-new application, Cap Gemini suggests looking at the following architectural questions:
- Should I use containers or serverless?
- How can I ensure my application runs reliably and cost-effectively?
- What is the best storage/database technology to use?
- How can I best ensure data privacy and compliance?
Of course, don’t forget technology needs experts to manage and run. So your architectural design needs to make room for acquiring or upskilling talent or working with tech partners who can share their skills.
Throw out the old economics books
Moving away from a capital expenditure model requires a very different metric from long-term demand-planning, outlays and depreciation. Simply supplanting average cost and infrastructure-utilization rates with cloud-based numbers will not help.
Good cloud economics need you to evaluate capacity demand at any given moment. That means your cost efficiency is tied to how well you can predict your capacity needs at the time of need so that you won’t be paying for unused capacity.
Instead, McKinsey believes “companies instead need to develop a dynamic operating-expenditure approach to cloud economics that continuously optimizes incremental costs by choosing the cloud services that best match their current workload requirements.”
FinOps is one approach. It fills the current gap in visibility for application owners so that they can link “business drivers of their cloud spend and the corresponding impact of cloud spend on unit economics.”
Another practice that is gaining ground is cloud observability. It offers better visibility into applications and the underlying architecture, especially in large distributed Kubernetes-based environments.
Get a new playbook — today
Cloud economics needs a mind shift. But this can’t be done overnight after years of optimizing IT capital expenditure.
Clearly, the cloud will be part of every enterprise’s future. And it is also going to get more nuanced and complex as companies try to compete with edge processing and other new architectures.
The sooner you throw out that old playbook and refresh your approach to cloud economics, the better it will be for your company as it competes in a cloud-driven playing field.
Winston Thomas is the editor-in-chief of CDOTrends and DigitalWorkforceTrends. He’s a singularity believer, a blockchain enthusiast, and believes we already live in a metaverse. You can reach him at [email protected].
Image credit: iStockphoto/Nuthawut Somsuk
Winston Thomas
Winston Thomas is the editor-in-chief of CDOTrends. He likes to piece together the weird and wondering tech puzzle for readers and identify groundbreaking business models led by tech while waiting for the singularity.