This is a follow up to a blog series where we explore DIY cloud cost management, monitoring, and optimization tools. In the first, we investigated how Lyft, Netflix, Segment, Expedia, and Slack built theirs. Check it out here.
“Build versus buy” is the eternal question for every engineering and DevOps team setting out to solve a technical problem.
In the cloud cost monitoring and optimization world, both options can take on many forms. For example, you can entirely DIY it — and build your own from scratch. Or, you could use free tools like AWS Cost Explorer and customize them. Either way, when you’re making the choice between a customized solution or a SaaS tool, there’s a lot to consider.
We often get asked about why companies should consider purchasing a tool, rather than building something homegrown. So, we put together some things to consider as you decide which approach to take.
Why Companies Build Their Own Cost Tools
There are a lot of reasons an engineering team may choose to build their own tool, but one of the trends we’ve seen, especially when we talk to cloud-savvy companies is that they feel like the popular cost optimization companies, who have been bought by larger, slower companies — have slowed their innovation.
As a result, they aren’t keeping up with next-generation services, like Kubernetes or Serverless. They were built for a time when optimization meant managing RIs and rightsizing — rather than architecting for cost optimized infrastructure. And while they are trying to support sprawling multi-cloud infrastructure, cloud-native companies are increasingly taking a single cloud approach, while building with best of breed technologies.
As a result, top tier companies like Netflix or Lyft, who have the highest tier (and most expensive) engineers have built their own.
(We’ve written extensively about that here.)
These companies don’t just want to know which RI plans to buy. They want to understand how cost maps to their business — and whether their unit economics are healthy. Because they didn’t believe a cost optimization tool that met their needs existed, they built their own.
So, should you do what Netflix and Lyft did? Here are a few questions you should ask yourself.
Question 1: How Many Engineers and Hours Will This Require — and How Much Will It Cost?
It’s no secret that engineers are in short supply — especially top tier ones. For companies like Netflix or Lyft, they have entire teams of 5-10 engineers who build and maintain their cost tooling.
So, before you jump into a DIY tool, scope out what it would take in terms of resources to build your own. Then, multiply the number of engineers you’d need by their average annual salary.
That will give you a complete understanding of the cost, plus the opportunity cost of building your own tool.
Question 2: Is This Initiative a Strategic Part of Your Business?
The second question you need to ask is how this will deliver value to your customers and differentiate you against your competitors. It’s not just about paying your engineers — it’s about the opportunity cost having them not building other features. Before you dive into a DIY cost tool initiative, have a conversation with your product management team about where it fits in your priority list — and whether this should outweigh other items on your roadmap.
If you still have doubts, consider this: DevOps experts Nicole Forsgren, Gene Kim, and Jez Humble dive into this topic in Accelerate, a study of why certain technology organizations are high performing while others aren’t. They found that high-performing organizations focus internal teams on core strategic aspects of their business, while software that is not strategic should in many cases be acquired using the software-as-a-service-model.
Building anything that isn’t a strategic part of your business can distract from your main mission — preventing you from achieving your full potential.
Question 3: Why Are You Better Suited To Solve This Problem Than a Company That Dedicates a Full-Time Set of Engineers To It?
Ok, here comes the part where we talk about ourselves. There are companies out there (and we’re of course one of them) with full-time engineering and product teams, dedicated to solving cost problems for our customers.
There are over 175 different AWS services, not to mention many flavors of containers and Kubernetes, and other services, like Snowflake that your engineering team is using — and all of those require deciphering and support.
Our mission is to think about how to solve important cost problems for our customers — and we think about it all day, every day.
Are you ready to dedicate your engineers to do the same?
Check Us Out
If you’re considering building your own homegrown cloud cost management tool, check out CloudZero. We’re built to empower engineering teams to control cloud costs by connecting technical decisions to business results.
The Modern Guide To Managing Cloud Costs
Traditional cost management is broken. Here's how to fix it.
