Summary & Key Insight
Interest in public Cloud infrastructure is surging. Across all industries and geographies, we see significant demand for insights, adoption patterns and guidance focused on pricing and sourcing public Cloud services. This represents a dramatic shift: just 24 months ago, a majority of ISG clients were focused on on-premises private clouds, or virtual private clouds hosted by large, Tier One and Tier Two outsourcing providers.
However, public Cloud continues to be viewed by many clients as simply a lower cost alternative to traditional outsourcing. This is not always the case. Public Cloud providers leverage a fundamentally different operating model than traditional data center outsourcing providers, and therefore, have a fundamentally different commercial structure. In this Leadership Insight we analyze the public Cloud operating model, and how it leads to a new pricing paradigms, which we represent as the Four Pillars of Public Cloud Pricing.
As sourcing professionals, we are inundated with news about successful transformation using public cloud—the majority of which is focused on how much companies are saving using services like Amazon Web Services, Microsoft Azure and IBM SoftLayer. This puts tremendous pressure on IT and sourcing organizations to adopt one or more of these services, oftentimes without a coherent Cloud strategy, or even a framework to measure total cost of ownership (TCO).
Public Cloud can be a transformative tool—enabling developers to get products to your clients faster and helping infrastructure leaders create a platform to dynamically expand or shrink the data center on-demand. However, as with any technology investment, it needs a business case. And the business case for public Cloud looks very different from that for most traditional IT and traditional outsourcing. This is because public Cloud infrastructure providers operate differently than traditional outsourcing and hosting providers: they standardize to increase volume, rather than a more traditional model of customizing to increase margin.
Given the fact that many ISG clients initially approached public Cloud use as a value-for-money challenge, attempting to use public Cloud as a lever to reduce costs, we created the ISG Cloud Comparison IndexTM, which focuses on comparing the costs of public Cloud to other more traditional delivery models. Based on this research, we developed the Four Pillars of Public Cloud Pricing, represented below:
- Pillar #1: Cloud is not always cheaper. Unlike traditional IT where assets and labor are the key drivers of cost, usage is the key driver of cost in the public cloud. High levels of cloud usage can create scenarios in which an on-premises option is more cost effective conversely, low levels of Cloud instance usage can create a scenario in which on-premises is more expensive. This is especially the case when an application uses features that are priced based on usage, such as such as data transmission and data access. This usage- based approach creates challenges for IT organizations because quantifying usage (compute cycles, disk I/O, bandwidth usage, etc.) is rarely done—if it is, it often does not match how public Cloud providers price, creating a lengthy normalization effort.
- Pillar #2: Public Cloud creates significant cost avoidance opportunities for volatile workloads. Sourcing leaders should avoid viewing public Cloud only as a lever to reduce operational costs. That said, using public Cloud for volatile workloads can be used a lever to drive substantial savings as a byproduct of the public Cloud operating model. For example, if a workload requires 25% more computing power one time per month, it can be very attractive to use public Cloud to accommodate this spike, rather than using capital to fund the incremental 25% of infrastructure needed to accommodate the spike, given that these incremental assets may go underutilized for the balance of the month. The effort to build, test and maintain this hybrid Cloud architecture should be accounted for in the business case, but if the incremental 25% of computing power is significant, the cost avoidance opportunity can be substantial.
- Pillar #3: Applications that take advantage of automation to de-provision unused images can help clients capture potential savings. Since usage drives cost in the public cloud, applications with the most variable usage patterns are strong candidate to move to the public cloud. However, if the images that are needed to support these volatile workloads are not shut down when not in use, the entire business case breaks down. Therefore, building this automation into applications, or using third party tools to drive this automation is critical. Again, this requires additional investment in development and or tooling, so should be accounted for in the business case.
- Pillar #4: Mapping usage patterns and technology requirements to public Cloud providers will help enterprises estimate the lowest-cost option. Combining the fact that many organizations cannot easily quantify usage patterns of applications with the variable nature of pricing across IaaS providers, buyers won’t know the full cost of the application until it’s actually running in production. This creates a chicken-and-egg situation for many sourcing and finance professionals as they work with IT on cloud-related projects. To secure funding for a project, a business case is needed; but the business case can’t be created until the application is actually running. There are two approaches to around this dilemma. First, to get an indicative price, use the online calculators each public Cloud provider offers. Cloud aggregation and normalization platforms like Gravitant or CloudGenera will make this process much easier and faster (Note that online calculators will only tell you the monthly run cost of your application, a TCO analysis should include a much broader set of criteria). Second, build a prototype. Having a smaller version of the application, or components of the application, consume public Cloud services prior to committing to building the full version can help you identify areas that were potentially not accounted for in your business case.
ISG clients are increasingly basing sourcing decisions based on applications—both individual applications at a line of business (LOB) level, and bundles of applications at enterprise level. Using a supply chain approach to evaluating, buying and managing Cloud is beginning to slowly take shape (1710SST, The Cloud Multi-Sourcing Framework: Democratize Access, Increase Control, Jan2016), and public Cloud will increasingly be a choice for developers and LOB users within this supply chain, competing alongside internally developed and outsourced private clouds as a business service that can be evaluated and purchased via a dynamic portal. Therefore, sourcing leaders need a strong understanding of not only how public Cloud providers price, but also the cost advantages inherent in using public cloud.
When developing a business case for public cloud, it’s important to note that there are a significant number of additional costs, as well as potential “soft” opportunities related to public cloud. For example, new tooling, new processes and new skills will be needed, none of which will be addressed by using public cloud calculators. On the benefits side, agility is a major factor in why many organizations are moving workloads to the public cloud. For example, if an organization can move to a continuous deployment model because they can automate the deployment of infrastructure faster, this can add an entirely new lens to the business case—one focused on opportunity, rather than simply cost. This is outside the scope of this Leadership Insight, and will be addressed in a future report, but keep in mind that given that this is a soft cost, it will difficult to quantify. This is one of the primary reasons why we see public cloud being used as a platform to build and deploy new applications, rather than legacy applications.
Enterprise Leadership Actions: When evaluating a public cloud, remember that the disposition and consumption patterns of the application(s) will drive price, which in some instances, could be more than on premises on a run rate basis, but may help to avoid significant future capital expenditures. Focus on the “Q” side of the price times quantity (P x Q) equation when developing a business case in the public cloud. While price is negotiable as commercial term (SST1740, Cloud Procurement: Focus on Commercial Terms, Avoid the Operating Model, Feb2016), more value will be found by focusing on quantity. Public Cloud vendors are incented by volume. Therefore, unless enterprises are a consuming significant volume of public cloud services, price discounts are likely to be negligible.
Provider Leadership Actions: Proactively help clients gain value from the public cloud, regardless if it’s a proprietary cloud, or a brokered cloud. Help them optimize the instances they are using for existing workloads. Help them decommission instances that are not in use. This is contrary to thinking for many providers, especially those in a broker role that gain revenue and margin from increased public Cloud usage. However, those providers that help client to optimize usage gain trust, and trust equals a relationship. As IT services become commoditized and standardized, trusted relationships will become even more important, and will become the hinge upon which sourcing decisions are made in the future.