Reviewing your cloud bill will only save you 25% of your cloud overspend. To unlock the rest of the savings you need to look deeper. Here are 2 other places you need to look to really optimise your cloud spend.
Every cloud environment has savings which can be spotted by looking at the billing and resource utilisation. Tools like Cost Explorer help highlight these so that you can make the required changes.
This may mean using reserved instances or savings plans, consolidating or simply turning off environments that aren’t used. In most clients I work with, there's a FinOps engineer who has the job of keeping on top of this, and who justifies his salary several times over.
But the truth is that these interventions are only the tip of the iceberg. For every million you save, there will be another 3 million that is missed. To save this money, you need to look at your application performance and your engineering practices.
Just because resources are 80% busy doesn’t mean that they are a good use of money. Capacity is just one aspect of performance - in all there are 7 that need to be considered:
Application efficiency has a direct impact on cloud costs. Some years ago I was assessing a new API for an e-commerce site. It was fast, it scaled - but it required 60 instances to run peak load. The engineers were asked to revisit the code, changed a line or two, and now it only required 6 instances. They’d immediately saved 90% of their hosting costs.
Each of those pillars of performance needs to be assessed in the same way. A system may be efficient to start with, but with costs which will rocket as it scales. It may be that instabilities require over-provisioning of resources.
And as with that API, there’s a second thing that needs to be looked at…
Cutting costs isn’t enough. To realise the savings you need to be sure they those costs will stay cut. And this means building a performance awareness across your engineering function.
This does mean making sure you have the right tools and environments in place. And the right data analysis around performance. But it’s much more than that.
It’s having the right roles and accountabilities so that everyone understands their responsibilities with regards to performance. It’s having the right processes in place so that all 7 pillars of performance are assessed across the lifecycle.
And it’s having the right big-picture business-level view of the impact of decisions. The team I was working with felt they had a good grasp of performance. What they had built was fast and scalable, just not efficient. But it’s little engineering decisions like that, which, over time, end up with a multi-million dollar bill.
Yes, reviewing your cloud bill is a great place to start. But if that’s where you stop as well, you’re missing most of what you could save.
Contact us and we'll figure out the full saving opportunity and guide you on the actions you need to take to save the money.