The trouble with TCO as it pertains to cloud computing is that the models require an understanding of the true value of cloud-based technology. This is much more complex than most enterprises realize. There are more moving parts that need to be assessed, as well as benefits that are there, but hard to define and determine. I propose we think differently about cloud computing and TCO, and indeed this is a core requirement as we move businesses to the cloud.
Which model to leverage?
Cloud providers have their own TCO calculation services that they use to show basic value, and they use this tool to sell cloud services. AWS’s is perhaps the most well known, and actually does a pretty good job of understanding some of the simple costs and benefits of using their cloud services.
But TCO calculation models that are for specific cloud provider do not go far enough to truly define the cost saving, which for enterprises is much more holistic. Indeed, if you just go by the tactical costs that cloud providers save you, you’re likely not getting the whole picture.
However, as we understand more about TCO and cloud computing, a few things become apparent:
First, the TCO of cloud computing is very much domain- or enterprise-dependent. The benefit you get from using cloud-based platforms, public and private, relates directly to the type of organization you are, and the business processes you support. Other factors should be considered as well, such as existing skills of the staff, existing investment in hardware, software, and facilities, as well as existing laws and regulatory pressures. Running TCO models that don’t consider the type of organization, as well as many other factors, are unlikely to result in a true holistic value of this technology.
Second, everyone is now discovering that the real benefit of cloud computing is not the cost savings it can bring, but the fact that IT can react much faster and more effectively to changes in the business. This is where TCO models morph from hard and easy to understand calculations, to grey areas that are difficult to understand and prove.
This is the value of agility and time-to-market, which is difficult to measure for those looking to leverage cloud. We know that there is value there, but the ability to measure that value in the context of a TCO calculation remains elusive. However, when I do the TCO analysis out in the corporate world, the outcome proves time and time again that this is the core value of using the cloud.
If it were just about cost savings and scaling, using public clouds would be more expensive, in many instances. However, the value of great agility and/or time-to-market would still make cloud computing a good move for the organization.
Finally, you have to consider your existing investment in hardware, software, and facilities. Many organizations ignore the millions spent on new supporting infrastructure, which actually becomes a key driver in defining the TCO, as well as defining a good timing to make the move.
Defining your TCO model
So, if TCO models are domain-dependent, then defining your own model will be part of the process. The best approach is to consider a list of attributes for the model, and then the degree to which these properties are relevant to your problem domain or enterprise. This will provide you with the ability to understand what benefits of cloud computing are most relevant to your business, and thus gain a true idea of the TCO.
These properties include, but are not limited to:
- Existing infrastructure in place. Those 10 million dollars you spent on a new data center, and servers to fill it, can’t be recovered. Leveraging cloud as a replacement for existing infrastructure means you have to factor the cost of sidelining the existing assets.
- Existing skills and humans. In moving to cloud-based platforms, there are new skills and talent required that are likely different from the skills and talent you have in place. How much will this part of the transition cost?
- Moreover, what’s the difference in costs from pre-cloud to post-cloud?
Cost of migration to cloud platforms. The amount of money it will take to move some applications and data to cloud-based platforms, including redoing aspects of the applications to better support cloud.
- Cost of cloud services when in operations. This includes the fees paid to the cloud provider, in the case of using public clouds, over a duration of time and should include a changing operational load.
- Value of agility (including time-to-market). Or, the value that the company places on the ability to shift or align quickly to the changing needs of the business. This should be considered around points of value delivered, such as the ability to quickly stand up cloud-based systems in support of a strategic acquisition, or to quickly scale up to higher storage capabilities in support of expanding R&D to new and emerging product lines.
- Value of avoiding future capital expenditures. The hardware and software budgets of the past get a complete makeover, with most traditional purchases being greatly reduced or eliminated altogether. This gets to the whole opex versus capex type of value.
- Cost of risk around compliance issues. If you’re in the healthcare or finance verticals, you’re already aware of the issues around compliance with rules and regulations. When moving to cloud, you could be adding or removing risk of non-compliance. These have costs to consider, and thus should be baked into the TCO calculation.
What’s important about doing this exercise is that you define something that’s more of a living model, versus something that just deals with an instance in time. As we become more aware of the value points we’ve defined in our TCO, and thus how to adjust them accordingly, the more accurate and valuable our TCO model becomes.
Most organizations don’t define the actual TCO, and end up missing the market when it comes to IT’s ability to deliver more value the business. If you go by cost data alone, you may save some operational expenses in the short term, but the justification of continuing the march toward cloud computing just won’t be there.