- SHARE
-
-
-
-
Here’s a dirty little secret in the tech world: companies use their own products less often than you think. Weird to think about, but it’s true, and something to laugh about at an employee reunion years down the line. That’s not the case with FinOpsly; use our own products, which have the direct benefit of saving us big money.
We’ve been bringing on large customers. And with those added workloads came huge spikes in Snowflake costs. Luckily our Snowflake optimization module went into beta just as we really needed to get a handle on these, and we decided to test it out on ourselves. We ended up being our own case study.
When we pointed it at our Snowflake environment, the results were immediate: a one-third reduction in costs, with zero impact on performance.
Truth is, Snowflake bills balloon for the same reasons everywhere. Warehouses got spun up, and we never turned them off. We oversized compute just to be safe. Redundant workloads had popped up, and we missed the overlap. Storage shot up far more than we expected. These aren’t the worst problems to have- obviously these happened because of increased business.
What changed for us was visibility. Once we broke our bill down by warehouse, workload, and team, the patterns jumped out. We could see which warehouses were running too hot, which were duplicating effort, and which were quietly sitting idle. Engineers had the data in front of them, so resizing and consolidating didn’t feel like a gamble. Workloads kept running smoothly, and the costs started sliding down.
Wins came from asking our product blunt questions: what warehouses are redundant? Can warehouses be rationalized? No drama, no downtime, just less waste. Accountability was a big driver here as well. Once costs were tied directly back to the teams using them, the behavior shifted. Business units could see their own consumption, and nobody wanted to carry the blame for unused resources. Add in pre-deployment estimates from Costix, and we now stop new waste before it ever hits the bill.
All told, a third of our Snowflake spend disappeared. More importantly, we built the guardrails to keep it that way, which is proactive FinOps: not waiting for overspend and scrambling after the fact, but preventing it from happening at all.
We don’t make dogfood, or champagne, or whatever else cliché that exists to describe using your own product. But knowing firsthand that your product can deliver exactly as your planned is a great feeling. And something that we can easily do for you, as well.
And shoutout to the original. https://www.cnet.com/tech/tech-industry/ellison-oracle-will-save-additional-billion/. Note the Oracle stock price at the time.
