According to Gartner, businesses will be spending about $333 billion by the end of 2022 on cloud infrastructure, and according to McKinsey, cloud spending will increase by 47% in the year 2021. These numbers are staggering and certainly depict a very positive picture here. However, cloud consumers need to assess the pay-off of such significant cloud spending. McKinsey reported that companies exceeded their cloud budget by 23% and that 30% of their outlays were wasted. This leads me to wonder if businesses have been able to optimize operations from their cloud investments. Whether the Cloud has just added to their costs or has it been good value for their money? And lastly, why some companies still grapple with mismanaged costs or added costs during their cloud journey?
These pertinent questions need to be debunked in times where companies are struggling to stay afloat and are trying to mitigate their overall costs. Cloud costs don’t necessarily mean IT costs but also include certain operational and managerial costs as well.
So, how do organizations harness the cloud cost optimization journey? Let me guide you through the same in this blog.
Why Cloud Cost Optimization?
According to Gartner, 45% of the organizations that perform a ‘lift and shift’ to cloud architecture endure higher costs and end up overspending by 70% in the first year. According to Mckinsey, “80% of the enterprises believe that managing cloud spend poses a challenge”. Flexera noted, “organizations waste an average of about 35% of their Cloud spend.
Other than just high overhead costs, poor cost management certainly reflects on business innovation and overall agility. Additionally, according to a cloud ability survey, more than 57% have experienced a negative business impact due to inefficient cloud cost management. This is because much of the importance is only given to cloud adoption and not cloud optimization. Organizations must look to save costs here and look to bring about a cultural and behavioral change to maintain a fiscal discipline. As we enter the post-COVID world and the next stage of the economic cycle, IT leaders must now work smart to ensure business efficiency through cloud cost management.
Six challenges in the Cloud Cost Optimization journey
Despite conceding to the benefits derived from cloud cost optimization, many organizations struggle with it. It is essential to address key challenges and hurdles faced by cloud users in optimizing cloud costs. Let me take you through some common ones:
- Provision for peak mindset: The easy access to point-and-click web consoles and APIs in the absence of capacity constraints can lead to “resource sprawl” and unexpected charges. This can be sometimes construed as a threat as companies will be unable to anticipate and assess how many resources they require vs. how many are being allocated by their providers. What do unused resources mean? More costs. Hence, provisioning can be a tricky factor here. To maximize the Cloud’s value, it is essential to apply the “just in time, pay for what you need” mindset.
- Ease of use and lack of governance model: Often, the inherent scalability, flexibility, and easy provisioning of cloud service can lead to resource sprawl and cost overruns. Lack of governance for cloud resources adds a multiplier effect to unexpected charges of resource sprawl.
- Complex, multilayered pricing and billing structures from hyperscalers: Cloud consumption bills are challenging to understand and make it difficult to build “budget vs. forecast vs. actual usage” comparisons. Major cloud platforms such as AWS, Google Cloud Platform (GCP), and Microsoft Azure do not provide standardization of billing models, billing formats, APIs, or services. Moreover, Cloud providers constantly change their billing and invoicing practices that can sometimes cause a complex issue. To make things worse, Public cloud pricing and billing structures are multilayered and tedious to understand. The lack of a standard billing model, formats, or APIs per the cloud vendors’ discretion adds to this complexity.
- Low visibility: Most of the customer has deployed solution for monitoring the Usage from Budget and financial engineering perspective. You can’t manage what you can’t measure, which often results in uncontrolled consumption.
- Too many options in cloud feature catalog: Complex cloud catalog options require careful consideration, and it’s not easy to find the best-suited feature with the lowest cost for a given context. Every year, cloud vendors announce hundreds of new services, features, instance types, pricing reductions, and even new pricing models. Organizations struggle to keep up with this pace and understand how each announcement affects their financials.
- Excess of alternative architectures: The same application can be built using many different architectures, services, and components resulting in very different costs. Organizations can struggle to calculate and identify the most cost-effective alternative to deliver their requirements.
Seven Mantras for Cloud Cost Optimization
Let’s investigate seven mantras that IT and Business leaders can use to accelerate their Cloud Cost Optimization journey.
- Cloud-First Mindset – Cloud deployment entails some structural and systemic changes in an organization. A cloud-first mindset helps organizations become agile in bringing forth these changes, whether in business or revenue models. It also helps if IT teams can make decisions around the movement of the Cloud based on the dynamic needs of various groups. Investing in PaaS capabilities and cloud-native toolsets can help here.
- Architect solutions for cloud economics – Cloud optimization is not something you do per se but rather a mindset you inculcate. An organization needs to arrive at the most cost-effective cloud architecture to meet their requirements by factoring in what’s on offer in the cloud catalog, including newer features, and knowing what resources to use by interpreting usage trends from billing.
In the past, organizations designed for availability, performance, and security to be delivered from a finite set of pre-resources planned for peak workload. The Cloud reverses this paradigm and allows for a more precise design that’s perfectly aligned to workload requirements. The architectural components in the Cloud carry a price tag, and thus optimal cloud architectures need to be designed with cost in mind.
Some of the core elements or principles of cloud economics today include:
a.) Unlocking the true potential of any cloud provider by accessing IaaS, SaaS, and PaaS.
b.) Ensuring cost transparency
c.) Right-sizing – During ‘lift and shift” migration to the Cloud, right-sizing is often overlooked or ignored. This may lead to oversized instances and unused resources, ultimately increasing the costs.
d.) Remove unused resources
e.) Workload Usage-based optimization using Autoscale up and down capabilities – as rightfully pointed out by McKinsey, organizations need to inculcate a “consumption approach” and need to “continuously match their demand with the best-fitting cloud services.”
f.) Aiming for Evergreen Cloud Managed Services (at least for Infra, Security, App Runtime) and always be on the lookout for advanced cloud features that be more cost-efficient than the previous one.
g.) Everything as Code – To maximize cloud economics benefits, establish a strong automation foundation with everything as code (infra, security, configuration, network, documentation).
h.) Automate your consumption forecasting and capacity management in the CI/CD pipeline.
3. Adopt a cloud cost optimization framework: Many organizations need to rethink what the Cloud can do for their business in the current climate. Acceleration and optimization of the Cloud are critical components to a successful cloud journey; both must be considered and intertwined. Whether an organization looks to optimize first for maximum cost and consumption efficiencies or accelerate first for greater scalability, there is no “best way.” Moving to the Cloud could reduce IT costs if it is planned and managed correctly. When you optimize as you go, the savings are significant, controlled, and scalable.
There are three key pillars to any cloud cost optimization approach:
a.) Optimize resources (rightsizing, right features)
b.) Create visibility and control (transparency of costs, Usage, and forecasts)
You cannot optimize Cloud cost if you don’t have visibility into the spent and a baseline. A good starting point for the Cloud Optimization framework is to ensure visibility of your spending and control over cloud expenditure.
Once you establish your requirements and budget, you must track and ensure visibility into your cloud spending. Once you deploy your application, maintaining control and visibility is vital to question the necessity of specific deployed resources and see whether they add value.
Organizing cloud costs could entail resource tagging, cost allocation, and chargeback and show back models.
Additionally, creating and using a clear BI dashboard for visibility and control can help your organizations tremendously in the following ways:
- It can help you to understand what you are paying for
- It will prompt you immediately when your resources are misused
- It will help you to keep cloud costs under control
c.) Establish effective governance (identify Usage, ownership, and department allocation)
To maintain an optimal state, you need to ensure that sound policies around budgeting are adhered to. In terms of Governance, the framework should oversee resource creation permissions as well. Microsoft offers automation tools like Microsoft Advisor and Microsoft Cost Management to monitor your spending and cost spikes.
Furthermore, organizations must look to reducing their monthly bills. Once you gain visibility into spending metrics, you must observe which unused resources can be disposed of and which resources could be optimized.
The journey for any cloud cost optimization starts with initial analyses of current cloud estate and identifying optimization opportunities across compute, network, storage, and other cloud-native features. Any cloud cost optimization framework needs to have a repository of cost levers with associated architecture and feature trade-offs. Businesses would need governance — the policies around budget adherence, resource creation permissions, etc. — to maintain an optimal state.
A practical cost optimization framework requires all three of the above. Achieving initial savings would entail analyzing the estate and identifying optimization opportunities across compute, storage, and networking, focusing on the highest costs first and/or incremental/additional cost, month over month- cloud vendors provide access to the costs and utilization. Businesses would need governance – the policies around budget adherence, resource creation permissions, etc., to maintain an optimal state.
The key here is to focus on quick wins first, followed by dashboard creation for better visibility and control. Lastly, establish a Governance model to maintain an optimal state.
To maintain an optimal state, you will need:
a.) Governance – the policies around budget adherence, resource creation permissions, and more
b.) Transparency – reports of costs, Usage, and forecasts
c.) Well define KPI
d.) Continuous Review
4. Continuous Cost Optimization Process:
The Cloud is ever-evolving, and hence, organizations must also ensure to evolve their portfolio as well. For example, automation, autoscaling, serverless services, containers, etc. have evolved the cloud game and, if adopted, can ensure continuity in reducing costs. Hence, it becomes of utmost importance to find new optimization strategies and opportunities to ensure continuous reviewing. The key here is not just to have a one-time cloud cost optimization journey but to ensure a continuous optimization cycle at every stage.
How often an organization optimizes Cloud depends on the Velocity of its cloud adoption, the speed at which it develops, and the alignment with its financial cycle. In general, given the dynamism around what is available and what is being used, cloud optimization must be a continuous process and part of an organization’s operating model.
Applying optimization practices right at the outset helps establish a culture of optimization and accountability. When thinking of optimization, while cost takeout is a sensible place to start, it’s crucial to think about the value (achieving a better cost to serve).
5. Implement a chargeback model:
Cloud consumers should be responsible for what they consume. Enable them to create forecasts and pursue optimization opportunities. A good starting point would be to develop a resource tagging (e.g., usage, ownership, department, and cost center) model to implement the chargeback model. With proper resource tagging, it is possible to associate resource cost to the resource owner — thus, a cost center code.
6. Use the right sourcing, pricing, and discounting model:
a.) Choose the suitable sourcing model from allocation-based and consumption-based services.
b.) Choose the right pricing and discounting models: Reserved instances are pricing options or discounts based on upfront payments and time commitment. Reserved instances can be purchased for one or three years, and hence, it becomes imperative to assess past Usage and make the right decision. Microsoft Azure (Azure Reserved VM instances) and Amazon (AWS Management Console) offer reserved instances and pricing models.
a.) Take a careful decision regarding RI as it works on a contractual basis.
b.) Always be on the lookout for potential price reductions and grab the opportunity before it ceases to exist.
7. Establish a cross-functional cloud FinOps team:
With the advent of pay-as-you-go (PAYG) models, financial decisions have been decentralized. This means that, in previous traditional IT models, only a few people were responsible for making financial decisions about infrastructure purchases. With new pricing models, anyone can make cloud spending and cost management decisions, and how has now become everyone’s responsibility. However, such decentralization also comes with added complexities. This may bring in inevitable bill shocks and deviations in budgets.
Hence, it becomes imperative to integrate FinOps. FinOps is nothing but a combination of FINance (budgeting and cost modes) and OPerations (infrastructure, apps, and data).
FinOps is a collaborative, data-driven way of managing Cloud spend that allows finance, IT, and the business to manage quickly. As McKinsey noted, most large enterprises will benefit by “bringing together technical, financial and sourcing talent into a cross-functional cloud financial operations (FinOps) team to manage cloud sourcing and consumption.” A good FinOps model combined with cloud sustainability benefits is about saving money and managing the Cloud’s use to make money.
Cloud cost optimization calls for a paradigm shift at the organizational level and at the behavioral level to ensure that cloud investments are utilized responsibly and optimally.
Cloud cost optimization is not just an operational concern or merely about “cost reduction”; it’s a value-driven strategic move. The path toward it will not be linear and requires tight collaboration among governance, architecture, operations, product management, finance, and application development to be successful. Cloud services are promising and futuristic, and with the right cost optimization strategies, organizations can not just increase their ROI and TCO. However, they can also put their businesses in the best shape! As DevOps has revolutionized the development processes, a good cloud cost optimization framework and FinOps can help businesses realize the Cloud’s real business value.
With the right strategic interventions, control, and operating model, the Cloud provides excellent visibility to organizations on IT spends and is undoubtedly the most crucial and promising/futuristic technology investment an organization can make.