It’s almost impossible to create an optimized cloud system out of the gate and keep it running at a perfect balance over the long term. There’s almost always something that could benefit from some tweaks and adjustments.
Cloud costs have a way of creeping up slowly while you’re not paying attention. And if you’re not careful, they can spiral out of control before you realize anything is happening. A new launch may churn through resources at an unexpected rate, or a new mega-client might dial up inefficient, expensive resources previously rarely used.
Unless you have a system to monitor your cloud costs, you might not notice these things happening — or know how to fix them if you see your costs climbing.
That’s what cloud cost optimization is all about — identifying these areas that need improvement and then determining which problems should be fixed and in what order of priority.
You won’t ever achieve a fully optimized cloud. Cloud environments tend to be so large and complex that conditions will always be in flux. The nature of constant change means you’ll never have a cloud environment in perfect homeostasis.
It can be helpful to take the mindset early on that you should always think about finding cost optimization opportunities and judging if, when, and how to act on them. Cost optimization isn’t something you do every once in a while and then forget about for long periods.
You never know when changing conditions might present a new opportunity to improve margins that weren’t there yesterday.
On the other hand, just because you spotted a great opportunity today doesn’t mean you’ll be able to easily find it tomorrow or even that it will still be a worthwhile change to make if you do find it.
With that in mind, let’s talk about identifying these opportunities in the first place.
How Do You Spot Cost Inefficiencies Needing Improvement?
1. Use your expertise and experience
The first step is to look at your cloud costs, not based on your infrastructure but based on what your business needs and what it offers to customers. It’s easy to see your cloud bill and that you’re spending certain amounts on storage, machines, or networking, but it’s difficult to drive change with only that information.
Instead, translate that bill into something that makes sense for your business. For many companies, breaking costs down by products and features makes the most sense.
That way, you’re no longer focused on the fact that you’ve spent a given amount for 500,000 machines. You can look at how you use those cloud resources.
If you have spent the past three years building and selling a particular product to your customers, you and your coworkers have plenty of working knowledge about that product and its features. This knowledge makes it much easier to put your expertise toward optimization.
Perhaps you’ve just launched a new product, and your costs have skyrocketed. If the changes happened simultaneously with the release, it’s a good bet that your new product costs proportionally more to support than your other products.
This is a red flag signaling indicating the new product has at least one significant inefficiency. However, you can’t judge how a product launch is doing unless you can compare it against other, similar products.
That’s just one example of how it’s much easier to identify inefficiency areas by organizing your spending based on business context rather than infrastructure.
2. Track patterns in your changing margins
This method of detecting cost optimization opportunities requires you to be able to break your costs down into unit economics. To do this, you’ll need a specialized platform — such as CloudZero — that allows you to track costs related to each product, feature, customer, or other categories of your choosing.
Once you’ve tracked measurements of unit economics, such as costs per product and costs per customer, for a while, you’ll start to notice useful patterns.
Let’s say one product costs $1 million per month, supports 100 million users, and generates significant revenue for your company. A second product costs $1 million per month, supports 1,000 users, and doesn’t generate enough revenue to outweigh its costs.
These two products may have appeared equal when you were simply looking at total costs, but as soon as you dig down into the costs per customer and costs compared to revenue, these products are not even close to being equal in terms of value to your business.
The more you can understand each product and feature down to the margin level, the more easily you will notice when something isn’t working well and should be optimized or rebuilt.
How Can You Gauge Whether It’s Worth Acting On An Optimization Opportunity?
Let’s assume you’ve used the methods above and pinpointed at least a handful of ways you think your cloud products could be improved. You might be tempted to work through these issues and solve them one by one, but this isn’t always necessary or helpful.
Since you’ll never have a fully-optimized cloud environment, deciding whether your cloud leans toward the more or less efficient side of the spectrum comes down to a judgment call.
As you plan your optimizations, ask yourself the following questions to determine whether an inefficiency is worth the time and effort to fix:
- Is the inefficiency significant enough that it’s wasting money?
- Does the wasted money constitute a reasonable percentage of your budget?
- Is the issue identifiable based on your knowledge about your products or did it become apparent when you calculate your unit economics?
If you can narrow your list of inefficiencies down to the ones that check all of these boxes, you’ll have a solid place to begin your optimization efforts.
It’s a better idea to focus only on the efforts that will be the most impactful for the least amount of wasted time and resources.
If you’ve identified a dramatic inefficiency related to an infrequently-used product that you’re not spending much of your monthly budget on, you might be better served by prioritizing other optimization opportunities first.
In fact, even after you’ve addressed more pressing issues, it might never be worth your team’s time and effort to restructure an inefficient product without a cost impact. You likely wouldn’t want to spend three years restructuring a product to save ten dollars per month.
On the other hand, if two products cost $1 million per month, but only one brings in a profit while the other hemorrhages money, your bottom line is significantly affected by the inefficiency. This would likely be a good place to begin your optimization efforts.
Knowing Your Unit Economics Makes Cost Optimization Come Naturally
Finding the unit economics for everything your company offers isn’t merely one of several viable options for cost optimization. It’s the best way to identify and address issues with zero guesswork involved.
It also practically guarantees that optimization will always stay at the forefront of your company’s goals.
After all, if every employee is aware of how much a product or feature costs, how much revenue that product or feature brings in, and how much it costs to support different types of customers, it just makes sense to pursue further efforts in the areas that pay off again and again.
Your product team will know exactly which projects should take priority and your engineers will know how to build efficient, cost-effective software simply by having a clearer understanding of what works well and what doesn’t.
CloudZero can help you get there by providing all the tools and context you need to find your company’s unit economics.
Track the metrics that make the most sense for your situation, view changes over time, and even receive daily project budget updates straight to your Slack chats.
With this wealth of real-time information, you’ll naturally begin to see cost spikes and unexpected budget increases as opportunities for optimization.