The siren song of building a custom, internal cloud cost management platform is enticing. Many brilliant engineering teams are convinced they can come up with a bespoke solution that perfectly fits their needs. They look at their company’s unique infrastructure and decide they can DIY cost management without having to rely on an external vendor.
Believe me, I get the temptation.
As a founding engineer at CloudZero who has helped develop our product from day one, I want to issue a word of caution: Reconsider your plan to build unless you truly and completely understand what you’re in for.
As a founding engineer at CloudZero who has helped develop our product from day one, I want to issue a word of caution: Reconsider your plan to build unless you truly and completely understand what you’re in for.
The product we have on the market today is the result of years of intense, winding effort — and it requires constant upkeep to stay current in the quickly changing climates of cloud and AI.
If you’re considering building versus buying a cost management solution, I’d urge you to keep reading. I’m about to explain the circuitous, arduous journey we took to bring CloudZero to fruition. Our story may help you decide if the real cost of starting from scratch is worth the payoff.
Want to read another “build” story? CloudZero’s SVP of Engineering Bill Buckley — and also a founder — shares his own experience here.
Why Building Your Own Could Be A Waste Of Time
When we started CloudZero, I was employee number two or three.
I was there to witness the first venture capital funds arrive in the bank, which was certainly an exciting milestone. Any big project like this feels, in its early days, like a buffet of possibilities — and it’s so easy to look past how 99% of startups eventually fail.
Interestingly, we didn’t set out to build the cloud cost intelligence platform that CloudZero is today. We initially wanted to focus on chaos engineering, observability, and serverless technology. The goal was to engineer an architecture that would scale far beyond the typical requirements and would be capable of taking potentially billions of events per second and analyzing them in real time.
We hoped to be able to model complex systems and even virtually unplug servers to analyze the impact — a concept made popular by Netflix’s chaos engineering project. While that sounded fascinating at the time, we eventually realized that outside of a few extreme use cases, no company would risk unplugging their actual servers.
It’s like skydiving or popping a motorcycle wheelie: cool in your head but terrifying in real life.
However, this initial project gave us an unexpected advantage. We had an architecture built and ready for extreme scale. When our early partners kept telling us their biggest challenge was understanding their costs, we realized we could apply our platform in a novel way that could solve their visibility problems.
If an internal team starts today with the concrete goal of solving a targeted cost optimization issue, they may well build a solution for the initial problem. But they likely won’t have the fundamental architecture required to reach the same scale and depth as they could find with a purpose-built cost management platform.
We only lucked into that robust architecture because we had set out to solve a different problem entirely!
It took three years to find the right product-market fit
Even with the right architectural foundation, the journey to a viable product other people might want to use was still long and winding. I estimate it took us three years before we had a product we could truly stand behind in a wide release.
It took us three years before we had a product we could truly stand behind in a wide release.
During those first years, there was definitely some floundering going on. We earned some early revenue by taking checks for what was essentially consulting work, and we spent years learning enough about FinOps that we could inject those principles into our product.
We also met with some early partners who sent us off in terrible directions.
We’d spend time building exactly what they requested, only to realize the resulting feature wouldn’t be valuable to any other customers.
In other words, we took some lumps. And any other internal team trying to build their own solution would have to go through the same process of learning what works, what good reporting looks like, and what genuinely matters for board-level decisions. You don’t just wake up one day with these years of experience; you have to learn it the hard way.
Even teams with existing expertise should expect plenty of roadblocks. If your internal team manages to bypass the learning curve that caught us, they’ll still face two significant challenges: the sheer, overwhelming complexity of the FinOps domain and the risk of architectural failure.
FinOps can be mind-boggling to newcomers
FinOps is difficult because it’s so abstract.
Unlike building software for a physical problem — for instance a restaurant that needs a point-of-sale system — FinOps requires engineers to mentally hold onto complex concepts like the dynamic allocation of shared resources across product lines and not lose the thread as they work. It’s draining just to grasp these ideas, let alone to build something that illustrates them in real time!
Many internal teams will find themselves lagging behind industry best practices and scrambling to put together basic graphs to show executives cost data. Teams that already have FinOps knowledge would likely still take well over a year to build a solution with comparable depth and functionality to a readily available solution like CloudZero.
Teams that already have FinOps knowledge would likely still take well over a year to build a solution with comparable depth and functionality to a readily available solution like CloudZero.
Development hiccups will be inevitable
Rushed development kills architecture. When small teams hurry to build bespoke tools that satisfy urgent demands from executives, it’s almost a guarantee they’ll find the limits of their system before they hit their goals. We discovered this the hard way early on; we started with a handful of EC2 machines and quickly ran out of bandwidth.
We pivoted to an almost entirely serverless model, but our recovery was the exception rather than the rule. An internal team may not even realize they’ve hit an architectural limit until it’s too late. They’re then forced into a massive, expensive architecture overhaul instead of delivering actual value in the eyes of their executive suite.
Worse, if an internal team builds a hyper-specific tool that fits their current development environment, what happens a couple years down the road when the company switches to new architecture?
If an internal team builds a hyper-specific tool that fits their current development environment, what happens a couple years down the road when the company switches to new architecture?
The team would have to learn the new platform and rebuild the bespoke tool to work with the new setup. Realistically, you could be looking at multiple lengthy and expensive rebuilds every time something in your organization changes.
This kind of fragility is something we deliberately built the CloudZero platform to avoid. With just one tool, you can connect all your resources and service providers. And if you switch architecture on a whim, it’s no problem — CloudZero will work with your new system, too.
Let’s not forget the trust factor
CloudZero is built on a wealth of expertise that I don’t believe another company could replicate. Our founders brought experience from their diverse backgrounds in security, B2B, and B2C. Our CTO, Erik Peterson, has worked with the FinOps Foundation since its creation, sitting in on the first board meetings and meetups.
Today, our platform gathers data and generates insights for over 250 customers. We’ve seen how FinOps is best executed across numerous industries, sales models, and every conceivable architecture combination. Our financial account managers (FAMs) know how to help you manage costs whether you use a few massive machines or thousands of containers. We can input costs from one cloud account, from thousands of cloud accounts, or from environments spanning multiple cloud providers and SaaS products.
In contrast, when you build your own platform, you only gain experience within the confines of your own company. You may not even realize there’s a better way to do something if you haven’t seen dozens of other companies dealing with the same issue. And if you’re looking for investors, limiting your knowledge base can only hurt your chances of securing funds.
Secondly, our platform manages costs for high-profile clients including Coinbase, Toyota, Duolingo, and Grammarly, among others. We’re SOC 1 and SOC 2 certified. I mention this because, at the end of the day, our clients have to trust us. They use our data to report financial numbers to federal bodies up to and including the Federal Trade Commission.
Imagine leading an internal team that’s been asked to generate a financial report for your company. Who’s going to double-check your math and make sure your calculations are precise and accurate?
Imagine leading an internal team that’s been asked to generate a financial report for your company. Who’s going to double-check your math and make sure your calculations are precise and accurate?
We’ve all heard the horror stories (Volkswagen comes to mind) where developers were told to “fudge a number” and wound up facing criminal charges. The financial and legal risk of generating reports is genuinely scary! And your internal team might not have the time to develop a comprehensive test suite that confirms every calculation is correct. By using CloudZero, you offload that legal liability, and your potential mistakes don’t have to keep you awake at night.
Finally, CloudZero provides access to valuable relationships through our extensive partnerships with companies like ProsperOps. Not every company has the bank account size to yell at Amazon or tell Google where to stick their data transfer fees. We can provide the necessary weight, leverage, and partnerships to help you out when it’s time to negotiate a better deal or solve a problem.
Let Our FAMs Be Your Guide
Let’s say you successfully build a custom cost platform. You jump in, engineer your DIY solution, and you say, “OK, what’s the big deal? What was this guy on about?”
Your company’s next problem is getting stakeholders to actually use the thing you just built.
If adoption is low, you just spent probably a year of your life and grew more than a few gray hairs for little to no reason.
At CloudZero, we have FAMs that welcome users to the platform, show them the ropes, and provide ongoing guidance and personalized suggestions for the duration of your contract with us. They do way more than just show you how to use our software; they’ll actually dive into your numbers in-depth to point out ways you could improve your finances. It’s a hand-holding experience our customers enjoy so much, some consider it a standalone service in addition to our main product.
Our FAMs can tell you what strategies work best for your specific industry, come up with ideas that might work, and stop you from wasting your time trying things that aren’t a great idea.
Without this guardian angel relationship, your company’s employees won’t know how to use the platform you just built. They won’t know where to look for the data they need, how to act on that data, or how to generate a report.
Without this guardian angel relationship, your company’s employees won’t know how to use the platform you just built. They won’t know where to look for the data they need, how to act on that data, or how to generate a report.
Guess who’s going to spend even more time explaining all those concepts to your C-suite?
Hope you have some free time on your schedule.
Buying A Solution Like CloudZero Allows You To Focus On Your Business
The blood, sweat, and tears that went into bringing our cost management platform to market were immeasurable. But we took this winding, multi-year journey so you don’t have to.
You always have the choice:
- Spend your next few years trying to reinvent the wheel, absorbing all the legal and financial liability in the process (not to mention locking yourself into a fragile and hyper-specific architecture).
- Or purchase a battle-tested, ready-to-use platform backed by years of expertise and absolve yourself of the frustration and potential legal headaches you’d otherwise have to face.
Want to see CloudZero in action before you decide? No problem! Schedule a free demo here.
The Cloud Cost Playbook
The step-by-step guide to cost maturity
