The cloud makes a big budget item for every modern organization in the post-pandemic world. But sometimes, its costs become too big for comfort. A billion-dollar private software company told Andreessen Horowitz that its public cloud spend made up 81% of the total cost of revenue (COR).
While the example above is an extreme case, it still reflects the struggle of many software companies.
Perhaps all the layoffs we’re currently seeing wouldn’t be happening to such an extent if they took care of their cloud spend first? In the example above, labor cost is not more than 19%, so cutting a mere quarter of the cloud bill would easily save more than letting their entire staff go.
Luckily, there’s a way out. But let’s examine the root of the problem first.
Why does the cloud come with a financial risk?
As their cloud usage grows and dominates IT budgets (expected to hit 51% by 2025), organizations scramble to lower their bills. Their first move is often to pre-purchase a specific amount and/or type of cloud services to benefit from discounts (reserved instances).
But when doing that, companies essentially assume many of their business variables won’t change during the reservation period.
This often doesn’t match the reality. Committed cloud spend is usually ~20% lower than actual spend, and some companies end up going over their initial estimations by 2x!
Stories from the industry confirm this. During one holiday season, Pinterest’s cloud spend went beyond the initial estimates by some $20 million.
So, even if you snatch a discount by committing to use cloud resources within a one- or three-year contract, you might not get enough of it.
Or the precious discounted capacity might get wasted when people forget to shut resources down when no longer needed or worse, run a job and forget about it. That happened to Adobe which racked up an unplanned cloud bill of over 500k because of a computing workload left running on Azure.
In either case, you end up paying for resources that don’t translate into any value for your business. Next thing you know, these hidden costs start adding up and become a significant financial risk.
3 biggest cloud cost risks
Lack of predictability
Imagine that you’re running an on-premises data center. If you buy a server, it comes with a fixed cost and specific operating expense. You can create your budget and stick to it.
But predicting costs gets tricky when you start using public cloud services like Amazon EC2 to run your applications. Your team can easily run 100 or 1000 servers at a given point in time, which is especially challenging because spinning up these new virtual machines takes a few clicks.
People might run servers for tasks and forget about them once the job is done. But you’ll keep paying for them, accumulating massive financial waste.
The complexity of pricing models
Most public cloud providers present their pricing across multiple pages and long tables. Services are charged using different metrics like time of use, data volumes, compute capacity, and data transfer.
And each price factor changes across dimensions like cloud region, instance type, or overall data volume.
On top of that, cloud bills are long, and all the cost data comes with a delay. You’re dealing with a lot of complexity here.
Many companies end up using only one cloud provider for multiple workloads, effectively locking themselves into the pricing of that provider. No issue, until a surprise price hike happens, like Google raising its pricing recently. And other providers might as well follow.
Vendor lock-in brings you many unwanted issues, from missed opportunities in alternative services to the risk of having the solution you’re using sunsetted.
The way out
If you want to future-proof your company against these risks, start by reducing your operating costs. Public cloud expenses are a great candidate for this because the market is full of solutions that can help slash your cloud bill without much extra effort.
By automating all the tasks related to managing and optimizing your cloud costs, you gain back control over your resources without having to hire additional engineers.
CAST AI analyzes, optimizes and keeps your cluster optimized to avoid cloud bill surprises. Book a demo to see how much you could save.
Leave a reply