When these five stages were first defined by Cloudability co-founder J.R. Storment, they were based on his extensive observations based on dealing with a wide variety of companies. Since that time, these stages have been further tested, codified and refined. Eventually, they became a core concept of the FinOps pillars:
With FinOps, these concepts have been honed into a truly workable cloud operating model, one that brings IT, finance and business together to achieve financial accountability in the variable spend of cloud. Find out more by reading FinOps: A New Approach to Cloud Financial Management.
With over $1B of cloud spending tracked in our system, we’re in a great position to see patterns in how companies are using AWS— and there are patterns. In fact, from tiny startups to experienced enterprises, we’ve seen again and again that the most efficient AWS users all tend to follow the same five-stage lifecycle.
We’re kicking off a series of blogs on this topic by introducing each of the Five Stages of AWS Cost Efficiency and the Cloudability tools that come in handy at each step. Take a look to see which aspect of the lifecycle your businesses is working on, and what your next steps might be— then stay tuned for an even deeper look into each part of the process in following weeks.
The first step in efficient cloud use is to make sure you have an understanding of ongoing changes to your spending. The goal of this stage is to create visibility into spending as it happens, and encourage accountability. This can be achieved through daily monitoring of your month-to-date costs, your rolling average, and estimates of what your end of the month spending will be. Read more about Stage I here.
You’ve got to understand where the dollars are going before you can optimize efficiency. This stage is all about splitting your spending by metrics such as Business Units, Products, Cost Centers, and Lines of Business. The first step is to put together a cost allocation strategy to determine what sort of metrics are relevant to your business; then, you have to implement a tagging policy to label the sources of your spending. The goal of this stage is to give each budget holder a view into their own spending, and to allow your finance and exec teams to view all of that spending, broken down by the taxonomy that best describes your business. Read more about Stage II here.
Don’t treat the cloud like it’s a datacenter. There are 168 hours in a week, and 108 of them are nights and weekends. All too often, companies ignore the downtime and over-provision resources, often leaving everything on, all the time. In this stage, the goal is to identify what resources can be turned off, sized down, or autoscaled back during non-peak hours. This phase is also about determining which instances you’ll need to buy reservations for, which you’ll do in Stage IV. Read more about Stage III here.
Start buying Reserved Instances iteratively and often. Companies tend to oversimplify their Reserved Instance purchasing strategy which results in cliffs and mismatches in RI inventory. Once you’ve removed waste in Stage III, you’ll want to run an efficiency analysis on your existing reservations and make any needed modifications to move them between AZs or within instance families. A small initial purchase to close any gaps you have is just the first step; you should really aim to revisit your reservations monthly or quarterly to continuously adjust your RI portfolio to match your infrastructure. Read more about Stage IV here.
Your cloud bill is going up and that may not be a bad thing if your business is growing. The only way to see if your efficiency efforts in stage III and IV had a real impact—and to get your exec team off your back—is to focus on Unit Cost as the most important metric to your cloud spend. Identify what business metrics you can marry to your cloud spending: cost per customer, cost per subscription, cost per stream, etc. Then marry these business metrics with your cloud spend data and focus your efforts on pushing the values of those metrics down, even if your total cost may be going up and to the right. Read more about Stage V here.
It’s important to note that this isn’t a one-time process that is ever truly finished. Your infrastructure will evolve, your footprint will grow. The very agility that makes the cloud compelling also requires diligence to ensure continuous efficiency. We’ve seen companies have the most success when they work this model as an iterative cycle into the way they manage their infrastructure. They rinse and repeat.