Cloud Cost Saving: A guide to choosing a rate optimization tool

Andy
15th January 2024
shutterstock 2080661035

Introduction

Cloud cost optimization is a complicated, but important area. This guide will help you evaluate cloud cost optimization tools and approaches that focus on AWS Reserved Instance and Savings Plan commitments – “Rate Optimization”. 

With the AWS pricing model offering commitments with up to 72% discounts, such rate optimization tools can provide great value in making your cloud budget go further. BUT, it’s important to remember that commitments carry risks as well as rewards. There is a lot of devil in the detail of how tools balance this risk and reward. Unfortunately, there is no common standard or easy way to compare them to help you pick the best tool for your environment. 

Without being a subject matter expert or spending large amounts of time researching, it is often hard to know what to look for. We’ll help you understand what questions to ask and how to challenge well-practiced sales and marketing messaging.


 

Key Takeaways:

  1. Ensure you understand how relevant this is to you. This affects the level of effort it is worth assigning and what integration with other tools, processes and people you’ll have to manage. All quite important as you consider whether to build your own approach or “buy” a tool to help.
  2. Rate optimization with AWS comes with risks AND rewards. It’s important to consider both when choosing an approach. Ensure you consider the risk of cloud or supplier lock-in and not just the savings.
  3. Different approaches will have a different impact on security, onboarding/offboarding and ongoing service delivery. Don’t assume they’re all the same and be clear that their specific impact is acceptable.
  4. Any cost optimization activity should provide a measurable return on investment and demonstrate its effectiveness. Ensure you have a clear way of measuring the value, that the supplier can be held accountable for that and you have an easy path to move away should they fail to live up to expectations.

How relevant is this to me?

Should Rate Optimization be a priority?

AWS cloud cost optimization is a very wide area. Beware of tools claiming to be a one-stop-shop for all your optimization and management needs. First establish what your priorities are and where you/your team’s time is best spent as that will guide what type of tool, if any, you should consider.

In the results of the State of FinOps by the FinOps Foundation (the leading community of Cloud Financial Management practitioners and thought leaders) observed that on average organizations are using 4.1 tools for cloud cost optimization and management. This reflects the challenge for any one tool to adequately act as a one-stop shop. 

Cloud cost optimization is such a wide area that a whole cloud financial operations, “FinOps”, discipline has emerged to describe it. It builds a lot of practices and tools around the basic calculation that cloud cost is determined by what you use and the rates you pay. To that end, AWS cloud cost management software tools tend to concentrate on one or more of:

  • Usage: identify ways to enhance the technical use of cloud services 
    These include finding efficiencies in deleting idle resources, scheduling resources to only run when needed, rightsizing to meet required performance, modernizing to use the latest technologies or harnessing cloud-native architectures.
  • Rate: using the cloud provider’s pricing model efficiently to access discounts
    This includes the use of AWS Savings Plan and Reserved Instances. These enable you to make a “commitment” to use or spend a defined level of particular services/resources for 1 or 3 years in exchange for a discount on that usage.
  • Visibility: surface cloud spend to measure, manage and drive accountability
    With so many cloud services available, all of which can be switched on and off very dynamically, often through developer’s code, it can be a real challenge to understand your cloud spend. However, understanding it is very important if your usage is going to be efficient and the rates you pay optimal.

Your first questions should therefore be:

  • What tools do I currently use?
  • What expertise do I have in my organization and how is it best used?
  • Where do I feel my immediate priorities for improvement lie?
  • Do I know how efficient we currently are so I can prioritize this properly?

The rest of this article focuses on rate optimization as that is our primary area of expertise. Rate optimization is a great area to engage a third party for help as the required strategies, data analysis and pricing model knowledge needed can be applied to any organization. Other areas of cost optimization often need more internal knowledge and as a result, can be a better focus of internal expertise.

Build or "buy"?

AWS Cost Explorer and Trusted Advisor are already available to you to provide insights and recommendations. You could use these to build your own commitment management approach instead of using a third-party solution. The challenges to overcome here come from choice and certainty, two issues that often prevent rate optimization initiatives from delivering value.

With up to 24 different ways to make new commitments and the need to manage how they interact with your existing commitments, deciding what to use when isn’t easy. You may find our Understanding AWS Commitments article useful in exploring this further. 

With commitment choices creating lock-in to either spend levels or specific service use you will also need a degree of certainty about your future cloud use. When you consider that the flexibility to evolve is often a key reason given for moving to the cloud, it then seems counterintuitive to take actions that limit that flexibility. 

Analyzing to understand where you could save and whether the saving is worth the compromise to flexibility is a real challenge. Are you in a position to do it? Do you know enough about the commitment types to build an approach that strikes the right, measurable and justifiable balance between discount and flexibility? 

How variable and predictable is your usage? Making commitments is easy if you always use the same resources, at largely the same levels and expect to for the long term. If this is you, then you might as well manage your own commitments. The more variability and uncertainty about the future, the more value there is in considering a specialist tool that can apply more advanced approaches.

When you consider the expertise needed, and the analysis, approval, purchasing and renewals processes involved it’s easy to see that even a relatively simplistic approach to commitment management takes a significant amount of time. Is that the best use of your, or your team’s, expertise? Could they add value elsewhere? If you reduce the time needed by taking a simplistic approach, how will you understand what savings you may be missing out on and whether the level of lock-in is a risk for the future? Given the scarcity of experienced FinOps practitioners can you afford the expense of recruitment (time, agency fees, onboarding)?

Every organization is different so only you can assess how much time you’re spending and whether it is worth the opportunity cost compared to focusing your efforts elsewhere. Did your organization really move to the cloud to become FinOps experts?

Evaluating Risk and Reward

Will I save more?

The market is full of eye-catching statistics about the amount you might be overspending on cloud and how much you could save overnight. Best-case scenarios and even averages often aren’t that helpful. You need to understand what it means for you, given your cloud usage and what optimization you’re already doing.

Look for tools that provide evidence-based estimates and then question:

  • What data do they use?
    • Does it cover a sample size that represents the typical extremes of your cloud usage so you can trust the results?
    • How do they make their estimates from your data?
    • Do they need access to potentially sensitive data or the ability to take actions that could compromise the delivery of your cloud service?
    • Can you review and control the access at a technical level independently or will you just have to trust sales and marketing material?
  • What insights will you get beyond the inevitable “you need to buy our solution”?
    • Will you learn how effective your optimization efforts are and how they compare to other organizations?
    • Will you learn anything about your current approach that you could use to improve it yourself even if you don’t buy from this supplier?
  • Are there other benefits beyond the financial savings?
    • Time can often be a more important resource than money. Will the learning curve or management of the tool cause you to swap one overhead or training need for another or will it release your time to focus elsewhere? What else could you be doing with that time and how does the return from that stack up against you performing the optimization activities internally?
    • Will it simplify or impede your current/future reporting and analysis processes?

You need to be able to assess all this quickly and easily so you can create a valuable shortlist that warrants reviewing the next level of detail. You may even choose to go and explore the market at this point and then return to this article when you have your shortlist to choose from.

For reference our approach to supporting this is described in our Free Savings Review and in our description of benchmarking and KPIs.

What is the lock-in/impact on flexibility?  

There are two forms of lock-in to consider. The lock-in to the optimization tool and the lock-in created by the commitments used to make savings. Ensure the distinction is clear and that you understand your options should things not work out as expected.

  • How does the tool use commitments to make savings? 

Different commitment types are better suited to certain use cases and come with their own options to manage the risk of lock-in (understanding AWS commitments). Are all these being used and how are decisions made on which to use?

  • What level of involvement do you want with the tool?

If you’re searching for a tool, maybe your preference is to let the tool manage everything on your behalf. If that is the case then make sure there is a way to audit its actions. At the very least you’ll want to know about actions that will create a lasting impact on your cloud spend should you stop using the tool. 

If you want the freedom to take a more active role to what extent is that enabled? Can you provide insights into future plans so that commitment decisions aren’t made purely on historical trends (AI and commitment management)? Can you exert any control over the nature of commitments being made for you so that the use of the tool doesn’t have to be an “all or nothing” decision? The low level of automation observed here points to the challenges of doing this and suggests a rightfully cautious approach to end-to-end Automation (State of FinOps by FinOps Foundation).

  • Buy-back and underutilization payments

AWS applies the benefits from commitments on a continuous, hourly basis through the term of that commitment. If there is no appropriate cloud usage to which the commitment can be applied in any hour, it becomes less effective as a way of providing overall savings. This is measured as “commitment utilization”. Utilization can decrease to the point where a commitment’s net effect is to waste, rather than save, money.

How will the solution maximize the value of commitments whilst avoiding underutilization and to what extent will you be locked into paying for it or wastage? This may be expressed in your contract termination rights so be sure to understand who actually “owns” any commitments. If there are buy-back guarantees or underutilization payments make sure you understand what is actually covered (all commitments, or quite often, just mistakes made by the tool), whether there is any cap and how your cash flow will be affected while any issues are resolved. The devil is always in the detail and this often makes reality very different to the sales and marketing messaging so be sure to explore this fully.

  • Contract term

Beyond the standard initial term and notice period considerations there are a few specific factors to think about. Can you terminate for performance-related reasons or convenience and how long would that take to complete? 

If you’re terminating for performance-related reasons you’re likely unhappy with the commitment management approach that has been adopted. How will you assess what your commitment position is and how will you take over its management? How will you independently assess performance on an ongoing basis (more below)? 

Be prepared to ask these questions upfront and push on the commitment strategy used (understanding AWS commitments). You may find the challenge of replacing an underperforming tool is so great that it, rather than the contract, provides the lock-in you’re aiming to avoid.

What’s in it for the supplier?   

Suppliers of AWS cost saving tools exist to help you save money but they must also make money for themselves. Ensure you understand what their incentives are as that will guide the way they operate. If your incentives are not aligned you may find you’re missing out on savings or being exposed to risks you’re unaware of.

  • Risk/Reward

Making any form of commitment comes with risks that can result in it costing, rather than saving you money. Who carries that risk is very important. 

Often suppliers charge a percentage of the savings they generate so if you carry all the commitment risk, or the supplier caps their risk exposure, it incentivizes them to use a more risky commitment strategy. 

On the other hand, if they are taking the risk they are likely mitigating that risk with higher fees or an overly cautious commitment approach. Either way, you’ll effectively be paying a premium in terms of the overall savings they deliver. 

Look for a transparent and balanced approach to risk and reward.

When exploring their approach it’s useful to understand how they combine:

  • Coverage (%): how much of your usage eligible for a commitment discount is likely to be discounted? The ideal is high coverage without the liability of lots of long remaining terms on your commitments.
  • Utilization (%): the benefit of commitments is applied on an hourly basis. Utilization is an expression of how many hours you had cloud usage to which the commitment could be applied. It may seem counterintuitive but the goal here is not necessarily 100%. Each commitment has a different tipping point in utilization where the overall effect switches from positive to negative. Striving only for 100% would mean missing out on some savings potential. It’s important to remember that “net-positive” utilization level for one commitment is not necessarily the same for another.
  • Wastage ($): is the cost associated with commitments where their utilization is so low that it actively costs you more than it saves. This should be minimized wherever possible.
  • Discount (%): what saving off the undiscounted (on-demand) rate is being applied? Are they promising high coverage but protecting themselves against under-utilization/wastage by offering you a low discount? In this way, they can make additional margin when times are good and limit/fund the downside when times are bad. If you’re not careful you can effectively end up paying twice for the service – one as a fee and one through this “insurance policy”.
    It’s important to consider all these metrics together to ensure the approach adopted maximizes the discount with a measured and well-managed level of risk. 

Satisfy yourself that high coverage/utilization promises are not delivered through large fees, small discounts or offloading all the risk to you.

  • Upselling

Is cost optimization just a route to upselling other services or an afterthought bolt-on to a wider set of services? If so, your rate optimization approach is not likely to be getting the appropriate level of attention and expertise. This will come at the expense of savings, undue exposure to risk or constant pestering from sales teams!

What Operational Impact is there?

What are the security implications?         

Ensure you understand what data is required for analysis and what permissions are needed to take optimization actions like buying, selling or changing commitments. To what extent can you independently audit and/or control this?

  • Data for analysis

Not all cost optimization tools use the same underlying data. This affects how light touch they can be in terms of initial setup, day-to-day operation and switching off should you stop using them. It can also, whether that is explicitly passed on or not, affect additional costs in delivering the optimization service. 

Ensure you understand what data is needed, the purpose it will serve and the frequency it will be collected. Then you can assess whether it is kept to a minimum that is capable of offering the scope of service you expect and need.

  • Permissions to act

If the cost optimizer will be actively managing commitments for you rather than just providing recommendations, what access do they need? Does this affect your AWS Accounts directly? Are you satisfied that the access they need is minimal and specific to the cost optimization actions you’d expect? What freedom do they have to change this and can you independently audit it?

What is the onboarding process?                 

Understanding the onboarding process is a great way to explore the implications of adopting a service. Not only in terms of the effort and people involved in completing it but also the potential learning curve, impact on your related processes and future offboarding if necessary.

  • What is involved at a technical level to bring your cloud accounts under management? 
    • How much of that requires input from you? 
    • Does it affect account ownership and if so how?
    • How would, and how quickly could, this be reversed? 
  • Are there any implications to your existing processes?
    • Are you still free to use all the same technical services? (e.g. some delivery models impact the ability to use “organizational level services”).
  • Can you continue to use all the same cost analysis or alerting tools? Some approaches create restrictions that limit the functionality you can access and/or force you to learn the supplier’s alternative.

Ensure you understand the tools used across your organization and explicitly raise them during evaluation. Unfortunately, this can often create a new sequence of functionality comparisons and onboarding effort as you adapt processes – work you’d have to repeat if you were to later offboard and no longer had access to the tool.

  • Will any new restrictions be placed on the activities of your internal teams or other suppliers? If so, what, why and how will they be enforced? 

For example, if the tool automates your commitment management it may prevent your teams from making commitments themselves.

Value Benchmarking and KPIs

How do I measure the value delivered?     

The value being delivered should be very transparent, both at a given point in time and as a historical trend. Ideally, you should be able to monitor this independently and not solely rely on the cost optimizer marking its own homework and telling you it’s doing a great job.

The financial savings generated should be easy to measure. Other benefits, like the time and headcount the tool saves you, will be harder, and more reliant on you. 

When defining your KPIs it’s important to consider what your original priorities were and the claims the supplier made about its impact. Having a view on KPIs and how you’ll measure them on an ongoing basis is useful when reviewing contracts to ensure you have a right to terminate if it does not live up to expectations.

Learn more in our article on benchmarking and KPIs.

Conclusion 

Cloud cost optimization involves a wide range of activities and expertise. Picking the right tools, delivered by the right people in the right areas is vital in taking a holistic approach. When done well it makes all the opportunities the cloud can bring far more accessible. Obtaining internal approval for projects and budgets will be easier and you’ll likely have released human resources to deliver those projects. 

We hope you found this article useful. We’d welcome the opportunity to be your guinea pig in using it as you evaluate cost optimization tools and would love to hear your feedback.

Learn more about our Automate Service.