Taking Control of Your Cloud Spend

Updated: Feb 23

In the past few years, we have witnessed our clients’ public cloud spend growing at a phenomenal pace. As a result, we often receive requests for guidance on ways to reduce the cost of the services being consumed. Clients often start with a small spend in the cloud primarily based on a demand consumption model, “pay as you consume”. Yet as consumption grows clients often begin gravitating towards contracts where committed annual spends in the cloud infrastructure are signed in exchange for steeper discounts. At first glance this approach appears to make sense: growing consumption is rewarded by the cloud supplier giving back with deeper discounts, ultimately lowering the unit costs of the services consumed.

In the event annual spend outpaces expectations, clients have discovered that the vendors are often willing to renegotiate larger discounts in exchange for larger annual commitments. When clients chase these discount hikes by allowing the vendors to track commitment increases closely to their forecasted spend increases, we say they are falling victim to the vendors’ strategy of “spend more to save more.” We’ve found that folks who try to play the game with the public cloud vendors are paying as much as a 40% premium over folks who rely on a disciplined approach and force the vendors to compete for their public cloud business.

How Cloud Vendors Force You Into Their Strategy Through Service Lock-In

Public cloud suppliers started out by providing a few core building blocks to developers, ways to quickly configure infrastructure to test and run their applications, as you can see in the graphic below.

Is IaaS a Commodity? An iceberg graphic comparison of core (compute, storage, network) to enhancements to core

Over time, the vendors began to provide enhancements to the core services. Many of these are available as standalone products from other companies, such as the analytics tools or content delivery services, but the public cloud vendors make their own proprietary tools available as convenient packages within their ecosystem.

In a broad sense, the core services shown above the waterline the tip of the iceberg are commodities. The services below provide a larger degree of lock-in to the vendor.

What can you do about it? Just as your organization would carefully choose any business critical software, the same rigor should be applied to any services that run major applications. It's very difficult and costly to unwind these workloads later, so they should be chosen deliberately, and carefully considered against the competition.

How to Inspect for Lock-In

To determine their level of lock-in, most customers start by examining a bill or other usage report. In many cases, it's easy to identify whether you're using a lot of core services, or services that contribute to lock-in.

Evaluating Core vs Lock-In in billing type, usage type, sw layer, hidden costs

This is a simplified version of a typical bill showing the service being consumed over the course of a month. It's not too difficult to separate those into the core and lock-in services and observe the level of lock-in in an environment. However, the bill may not contain the whole story because some lock-in services are invisible.

Here you see services such as AWS’ Big Data tool called EMR or one of their web development tools called Elastic Beanstalk, which have proprietary features that contribute to lock-in. They’re “free services” and you only pay for the underlying core services used to run them, so they may not appear on the bill. For example, you may not see that 12,000 of the 20,000 hours of the C4 compute instance are consumed by people using the Elastic Beanstalk service.

The bill or usage reports are best used as a starting point to pull the data together and figure out where to look deeper. The next step is to align with the various managers in the organization to understand what services are connected to which projects and applications. If not, you risk turning all the control over to the vendor and its “spend more to save more” strategy.

But out of convenience, customers are going right to their cloud suppliers for tools, and very often, that’s what limits their future options.

IaaS Loss of Control in tools, process, enhancements, people, contract, intellectual property

This graphic illustrates the slippery slope of cloud service spends, and how quickly one purchase turns into vendor entanglement at your business. This has been made even easier for the suppliers in the cloud because companies are going right to them to leverage their tools and enhanced services, often limiting future options they could be getting elsewhere, while often overlooking the impact on existing processes. This practice can have a devastating impact on cost and satisfaction of service.

So, how do you approach taking back control and optimizing that cloud spend by introducing competition?

IaaS Cost Optimization and Reduction - document spending, behavior analysis, value competition

It's all about gathering data to support the case for competition, and following a repeatable process, as outlined below. Remember the gathering of this intelligence is meant to give you leverage in determining what may be competed and drive a better cost and service value. The data can also provide intelligence to stakeholders of what might be considered low risk to compete and provide justification to the financial teams to support a compete decision.

Public Cloud RFP 3 phases - requirements definition documentation, rfp development, vendor selection and contracting

Because clients are usually under enormous pressure to move quickly to cloud solutions, the necessary due diligence step is too often ignored. This means organizations place little value on accurately defining requirements and using competition in a deal, which gives all the leverage to the supplier. To keep this from happening, we urge clients to take the time to conduct an effective RFP, and beware of the following vendor responses:

  • They ignore your guidelines on how to price service so you can easily compare them to others.

  • They incorporate inflated assumptions about your future capacity levels, and use them to justify buyer commitments to annualized spends.

  • They include services that are not yet released.

  • They include software and services not requested.

To make sure that the leverage is one your side of the table during your next cloud, or any IT deal, we recommend the proven process illustrated below.

5 Steps to Regain Control of Cloud Spend - understand playbook, understand motivations, strict governance, avoid lockin, create competition

The easiest place to start practicing this approach is with new cloud spends. You will see that these steps will hold your supplier to a higher standard, avoid automatic lock-in and preserve your leverage.

The following ClearEdge staff members contributed to this article: Analyst Jack Schneeweis, Director of Client Engagement Mike Martineau, and Managing Director of Cloud & Managed Services Jim Faletra.

This blog post was inspired by the webinar, IaaS: Competing for Savings. You can access the full recording to this webinar below. To learn more about this subject, view our AWS Executive Briefing or Microsoft Azure Executive Briefing, or contact your ClearEdge representative.