by Gianni Huaman and Patrick Ryan
When companies put together a business case for migrating to the public cloud, they typically focus on cost reduction and efficiencies. IT looks at comparable hardware, software, and personnel costs, on-premises versus cloud. But the primary reason to leverage the cloud should not necessarily be to save money, although that can be a happy byproduct. Rather, the cloud supports increased agility, flexibility, and scalability. A full reckoning of cloud-computing economics requires going beyond a Total Cost of Ownership mindset.
It is under-recognized that the capabilities that the cloud provides have been a critical enabler of the growth of Agile as a development discipline.
Firstly, your ability to operate your business at high speed and to be agile is enabled by the automation and capacity-on-demand nature of the cloud. It is very difficult to adopt modern, efficient business practices if it takes weeks to provision servers, or if the transition from small-scale development to production-scale deployment becomes a significant roadblock. But the benefits of being able to systematically operate in an agile fashion in terms of speed, cost, and quality of outcome can be very sizeable. It is under-recognized that the capabilities that the cloud provides have been a critical enabler of the growth of Agile as a development discipline, and, vice versa, that the needs of developers working with Agile processes has driven the arc of development of those cloud capabilities. They are, in many ways, two sides of the same coin.
Secondly, it is expensive to do high-end analytics with traditional infrastructure because the demands for storage and computing power can be extreme. Sizing data centers for the peaks is out of the reach of many companies, especially as newer analytical techniques are even more demanding of the underlying infrastructure. But the public cloud supports bursting capacity for short periods of time rather than permanently provisioning it. Take an example use case such as financial risk modeling (which uses large Monte Carlo simulations): it is only done periodically, but requires massive computing resources. For the cost of a single server, you could rent the equivalent of 1,000 servers for an hour every month to perform these simulations much more quickly. And the value of the time saved by your specialized employees probably outweighs the technology cost considerations anyway.
Finally, the public cloud also provides the ability to increase or decrease the scale of your infrastructure to suit market or business conditions; the option to abandon cloud investment if a project fails to achieve milestones or market demand does not materialize; the option to time the scaling of cloud infrastructure when market and/or business conditions are more certain. These scenarios all represent “real options” for the business and can, in principle, be valued the same way financial markets use option-pricing theory to measure the present value of the opportunity considering the investment required, the length of time for which investments can be made and/or deferred, the time value of money, and the riskiness of the project. Take the example of entering a new country (which may have specific data-localization rules). The risk profile is very different when it is no longer necessary to sink cost into building and operating data centers upfront.
Many companies are looking at the case for a wider move to the public cloud, but often from an overly narrowly perspective. A more complete assessment of the benefits of cloud computing can present a very different picture and may be the difference between a successful and an unsuccessful business case.