Public Cloud Buying Tips, Trends, and Traps

The cloud market today is still a two-horse race between AWS and Microsoft Azure, though AWS has notably lost some of its share to Microsoft and Google, as they continue to aggressively drive cloud adoption.


In terms of trends, AWS's savings plans are frequently coming up in client conversations, which are essentially a more flexible version of reserved instances. You still commit to a certain level of compute, for example, over a fixed term and automatically receive discounts off your monthly billing. They are slightly less restrictive than reserved instances regarding their applicability.


To prevent the loss of market share, AWS is showing new willingness to offer added incentives, such as credits, to compete for growth and new business, and is also more likely to respond to competitive requests, such as RFPs. The following chart shows these trends, and what’s happening at Microsoft and Google, as well.


Public Cloud Market View Trends and Observations, AWS Microsoft Google

Public cloud pricing is published and available on the provider’s web page, which typically shows hourly or monthly rates for a specific cloud service. Here is an example of what you will see:

Public Cloud Pricing AWS Azure Google

Customers can drive additional value by making sizeable multi-year commits. For example, AWS offers the Enterprise Discount Program (EDP), or Private Pricing addendums. Microsoft has cycled through several name iterations and offers “Commit to Consume”, and Google currently offers “Platform Commits”.


These agreements are typically structured as either 3- or 5-year deals, though on occasion, we have observed 1- and 2-year term agreements. In exchange for the commitment to spend a certain dollar amount on total cloud services over a given term, customers can expect to receive some level of negotiated baseline discount. (Some vendors call this a blanket discount, or their cross services, or universal discount.) But these discounts depend on two things: your contract length and the amount of dollars in your annual commit, illustrated below.


Public Cloud Annual Commits Deal Structure and Mechanics

AWS's discounting is the most predictable of the three cloud providers and tends to rely on a more established pricing structure to calculate their baseline discounts. But no matter which provider you choose, these discounts typically cap out between 20% and 29% for larger deals, such as $100 million or more per year in cloud spend. Deeper discounts are available from these providers when they know you are moving away from another supplier and want to accelerate adoption and consumption. However, it is important to note that discounts simply provide additional cloud purchasing power – they do not reduce your annual commit requirements.


As you can see in the chart, credits go by many names, but they are often applied to offset the cost of third-party professional services or specific cloud migration efforts. Regarding support, AWS appears to be the only supplier that contractually requires you to carry enterprise support.


Key Risks


We have listed some of the most common “gotchas” our clients face in public cloud deals.

Risks of Buying Public Cloud

The number one problem for customers buying public cloud services is overcommitting. This happens when customers fixate on maximizing their discount, and wind up overextending the organization on the forecasted commitment level in the process. But any benefit derived from additional percentage points in discounting becomes moot when they are forced to make massive shortfall payments to the cloud provider, which is why we always recommend taking a conservative approach to forecasting. (We have observed customers successfully re-open and renegotiate contracts when they greatly exceed their original committed spend.) Please note: Cloud providers will not allow you to reduce your commit retroactively: you use it or lose it.


Another major issue is the vendor’s access to critical information. Stakeholders who are in contact with sales reps often end up leaking your cloud preferences or demand for new technologies, which erodes deal uncertainty and shifts significant leverage towards the vendor. You can reduce the likelihood of this happening by issuing hush letters in anticipation of the deal, and by streamlining all communication through the deal team.


Other areas of risk in public cloud contracts include the renewal protections, portals where you can buy additional third-party software hosted on the provider platform, and support costs, as outlined above. It bears mentioning that Microsoft, AWS, and Google all have marketplace portals where you can conveniently purchase third-party software, but these prices can be much higher than those you could negotiate directly with the vendor. We have observed the “convenience factor” cost as much as 30% to 40% more.


Support costs present risk, and often go unnoticed when buyers negotiate their cloud spend. For this reason, we urge clients to push back on the sales rep if they fail to address the subject of support or try to delay the conversation until after the deal is done. These costs are usually calculated as a percentage of your cloud spend and can be assessed in advance.


The last major area of risk to clients is vendor lock-in caused by the lack of competition in a cloud deal. It is critical that you compete your workloads and consider a multi-cloud strategy wherever possible -- especially early in the supplier relationships so you can maintain options and leverage going forward. Failure to compete their workloads typically costs clients an average of 40%


Let’s review some of the best practices on how to compete workloads.


Reconcile Public Cloud Workloads, Compete

This graphic illustrates the process of competing workloads. The first step is identifying workloads. Loosely defined workloads are essentially any set of business requirements that could be hosted on a single cloud provider. Identifying and putting names to these workloads will greatly help in negotiations with cloud providers.


The next step is identifying which workloads are portable. There are a variety of technical or business reasons that a certain workload might need to be hosted on a certain provider, but some workloads are portable, and could be moved to any cloud vendor. These workloads can be competed and allow clients to create leverage with the vendors that will lead to more competitive discounting.


For a deeper dive into cloud purchasing best practices, read our blog titled Taking Control of Your Cloud Spend, or contact your ClearEdge representative.


- Jack Schneeweis and Kevin Cammarn are ClearEdge Analysts.