Our years of experience shows that organizations waste 30% of their hybrid IT spend, on average. This article identifies the five key components of a cost optimization strategy and how to be successful with each of them.
5 Key Strategies for IT Cost Optimization
- Align initiatives with business priorities
- Gain visibility across your hybrid IT
- Define cost structures
- Safely identify inefficiencies
- Allocate cost based on activity
What is Cost Optimization and How Do You Get Started?
IT exists to support critical business initiatives, but good IT service support comes with a price tag, no matter what size your organization is. There are multiple levers to pull in an organization to optimize the cost of IT support. Some are aimed at proactive prevention of wasteful behavior, while others are focused on adjusting or correcting decisions already contributing to cost.
There are two categories for these levers to pull to optimize the cost of IT support: proactive and corrective.
The best way to start an IT cost optimization initiative is to focus on the activities with the best and quickest return on investment. Be careful about setting your time horizon for payoff too far out. IT and the business will change too fast to be looking much more than a year or two into the future. Technology is constantly changing, businesses reorganize and merge with other organizations, and competitive forces and business strategies change.
Although there surely may be other sources of waste, the biggest savings potential can normally be found in infrastructure operations, whether that is resources you own in your on-prem data center or resources you subscribe to through a public cloud service. Every case of over-provisioned resources is a waste on its own. But it also has downstream effects on software licenses, system administration costs, and more. With the right tools, identification and right sizing of over-provisioned resources is a straightforward discipline that can even be automated.
In this article I have identified the most important components of a cost optimization strategy and what you need to consider when implementing them.
1. Align Initiatives with Business Priorities
The recommendations and best practices outlined in this article are about how to maximize the impact of your efforts from a monetary perspective—the core of cost optimization. But before you get into that, it’s important to reflect on other priorities too.
Rightsizing, auto-scaling, just-in-time provisioning and increased cost awareness are all activities that have a positive impact on the company’s results. By minimizing the spend, the bottom-line profit will increase. But it’s important to also consider IT’s objective to fuel top-line growth. In the era of digital transformation, IT is often the primary driver behind increasing market penetration, reducing lead times and spurring innovation. Only looking at the cost side of the equation may counteract the overarching business objectives of revenue growth. It’s important to always review optimization measures against such objectives. "Optimizing" cost means finding the right cost, not the lowest.
Another important aspect to consider is the importance of communicating to get buy-in from the business. Sharing the motives and rationale behind these initiatives is key. A lack of support for the changes you are planning can undermine the best-laid plan. Not consulting the business will reflect negatively on IT and miss the point that IT exists to support the business.
In your effort to change behavior to consider cost more, you need to be close to your customers—the business. It’s important to engage in regular conversations with business teams to understand their needs and plans. This will build mutual trust and buy-in on optimization plans. Moving away from a transactional relationship and interacting more frequently will also enable you to optimize your standardized offerings and minimize costly non-standard requests by shaping their infrastructure needs at an early stage.
Successful and lasting cost optimization requires mutual agreement on priorities.
2. Gain visibility across your hybrid IT
Optimizing the cost of IT services requires visibility of all the different components involved in the service delivery. But it’s actually pretty common for organizations to lack this visibility. Hybrid IT, involving both public cloud service and traditional on-prem data center operations, is the new normal for almost every organization. Service components across the board have been migrated to or launched using public cloud services. Many times, the placement was opportunistic (because of provisioning lead time, cost, security, etc.) and not part of a meticulous master plan. This has resulted in individual business services often spanning multiple platforms and service providers.
Along the same lines, the management tools being used were not selected based on a solid hybrid IT strategy either. They each cover a part of the environment or a specific technology and don’t offer a complete view. To understand the usage and performance of hybrid services, you need a tool that offers unified visibility and analysis for the complete environment.
Being limited by silos makes it difficult and time consuming to gather and analyze the data.
3. Define Cost Structures
To properly quantify cost savings, you need to define the cost structure for your operating environment. Breaking down the cost into individual components allows you to quickly valuate different alternatives and make sure that financial impact gets top priority. Optimization programs without continuous calculation of financial impact risk becoming simply a technical exercise and losing sight of the overarching goal: to reduce cost.
For cloud-based services (IaaS, PaaS, or SaaS) from a third party, this is a relatively straightforward process where the cost per instance, transaction, user, etc. is the foundation of the business agreement and subscription terms. The same applies to setups where hosting and operations of proprietary infrastructure have been outsourced to a third party. But in both cases there may still be some parameters in the agreement like volume-based discounts or term commitments that need to be reflected in the structure.
For traditional on-premise infrastructure, hosted in your own datacenter, you normally will have to do a bit more preparations. To calculate the cost of running a whole system or individual resources assigned to it, you need to define the cost of:
- Dedicated hardware components
- OS and hypervisor
- Maintenance and admin effort
These three are normally the big-ticket items that you should focus on first. For a more exact estimate, you could consider eventually adding cost for network connectivity via shared resource, facilities (power, cooling, floor space), and software licenses. But it’s important to not get stuck in trying to build the perfect structure. Start with a simple model that can be gradually refined by adding more layers.
Most organization have limited their options to a set of standard configurations that they offer. In those situations, the aggregated price based on the factors listed above can simply be predefined for each configuration.
4. Safely Identify Inefficiencies
Every organization strives to avoid infrastructure over-provisioning, but this proves easier said than done. The first challenge is estimating the optimal amount of resources to assign to new workloads and applications. No matter how much effort you put into it, you will likely be off by a certain percent.
In those cases where you get it right to start with, the conditions under which your applications are running will likely change as time goes by.
- Workload intensity will vary due to seasonality in the business
- Changes in business processes impacts the demand for services
- Resources allocated based on forecasts need to be adjusted to real outcomes
This means that you still need to evaluate and then optimize or even decommission over-allocated resources on an ongoing basis. In the process of identifying excess allocations, it’s important to include enough historical data to cover a full business cycle of each analyzed workload.
For resources allocated in public cloud, these right-sizing efforts will have a direct impact on your overall cloud expenditure. An automated right-sizing process should be part of your continuous management process and could considerably reduce your cloud cost.
For on-premise infrastructure, where the total cost is made up of a combination of Capital (CapEx) and Operational (OpEx) expenditures, the impact will be less direct and come with a slight delay. The CapEx component, the actual infrastructure investment, is sometimes considered sunk cost that can’t be recovered. But taking such a lazy approach is overly pessimistic. Your ability to recover the cost depends on if you can successfully repurpose the identified excess and replace or defer other planned investments. The reclamation and new demand for resources may not coincide, causing the bottom-line impact to be slightly delayed. But that should not be taken as an excuse for not trying. Most modern architectures based on hypervisors and pools of virtual resources offer good mechanisms for effective resource control and right sizing.
5. Allocate Cost Based on Activity
So far, we have focused on designing a process that corrects non-optimal provisions. By frequently analyzing allocated resources and determining corrective actions, you can considerably reduce the amount of waste.
But if there is no cost associated with the usage of IT resources by different business units in an enterprise, there’s a risk that each unit will continue to utilize the resources to maximize its own potential benefit, to the detriment of the enterprise as a whole. A simple way to further ensure effective use of resources is to charge each business unit in relation to their consumption.
For public cloud resources this is straight forward. Charging for exactly what you are using is a fundamental mechanism of the framework, and not relaying that cost to the business units means missing a great opportunity to increase cost awareness. Letting each cost center pay for what they use, while assisting them in optimizing the cost, will help drive the total cost down.
For traditional on-premise data center environments, it’s perhaps not as straightforward and will put some new requirements on your operational processes:
- You need instrumentation that accurately tracks activity by customer, especially to identify separate tenants using shared resources.
- You need to aggregate the total cost including deprecations of capital investments for the involved resources.
- You should also make a plan for overcoming cultural barriers towards internal charges, especially if your organization historically has considered IT an internal cost center.
If these requirements pose too much of a challenge and are severely impacting your timelines, look for compromises that can get the initiative off the ground quicker.
A well-designed allocation policy enables business transparency when it comes to cost and drives thoughtful business decisions for the whole company, requiring minimal allocation overhead and supervision once implemented. But it should not be your only measure. It will complement, but not replace, the continuous right-sizing efforts described in steps 2-4.
Rinse and Repeat
Once you have finished a cost optimization project, you will probably have identified a few things along the way that can be further optimized. IT cost optimization should be a continuous process of improvement. Chances are, your second go at cost optimization will be greatly simplified due to the realized benefits and trust earned through the success of your first time through.