Notes from the discussion on May 15th collected from member contributions covering - FinOps from Migration to Operation.
In this week’s Member Call, Ashley Hromatko showed us how Pearson integrates FinOps into their migration and operation process.
Before becoming FinOps Manager, Ashley Hromatko was a Technical Project Manager. When Pearson began moving their projects to the cloud, Technical Project Managers were doing a lot of work that was, in essence, FinOps. As their cloud grew, those FinOps roles become more important, and a FinOps team was born. Now Ashley is leading the FinOps charge, bridging the IT/Finance gap and proving the value of the cloud migration several times over.
This week, she walked us through their process from migration to operation. Here are three FinOps lessons that really stood out.
Lesson #1: Create a Gated Process
It’s a story we’ve seen time and time again: A company starts a cloud migration to become more efficient, dives in with both feet and runs into growing pains. Questions start arising about the value of the cloud, the return on investment, the amount of headcount needed and where the funding for projects is coming from.
Ashley’s solution is a multi-department gated process, but not one designed to slow or stop process. Instead, the gate is there to give projects the greatest possible chance of success. By taking the time to thoroughly plan out the financial, security and governance implications, they’re able to stop problems before they start. For that reason, it’s vital that the gate includes the right of refusal to halt problematic projects that are more cost-efficient on existing data center resources. By building a gated process, the FinOps team is able to ensure the health of the entire cloud, which is, after all, what FinOps is all about!
Lesson #2: Involve FinOps Early
A key step is to make sure FinOps is involved early in the process. We’ve all seen projects launch with little to no oversight, only to have FinOps come in after the fact trying to clean up messes and play catch up. Instead of constantly scrambling, Ashley’s team spends their time planning ahead. Sure, occasional issues may arise, like unexpected spending spikes or untagged resources from unknown cost sources, but they’re fewer and far between, which makes them easier to tackle.
And when we say early, we mean early, like during the initial migration proposal phase. The whole point of FinOps is to bring people together to avoid surprises, maximize the use of cloud spend and better achieve cloud goals. Getting involved in the early planning stages allows the FinOps team to make accurate estimates of potential costs and make sure everyone’s on the same page — and that the budget is there before it starts being spent.
Lesson #3: Use Budget Alerts Against Estimates
Many vendors and third-party applications offer alert services when spend goes above a pre-set limit. While these are a great way to stop out-of-control costs, Ashley showed how they can be taken to the next level when you have dependable estimates from your FinOps team. By setting the budget limits to coincide with the estimates, alerts can be used as a warning system to help development teams stay within their estimated budget. (It also helps to build cost awareness within engineering teams.)
However, there’s an extra layer to this approach that meshes particularly well with FinOps. Instead of reacting to the alerts in a punitive manner, Ashley recommends responding with a spirit of camaraderie. Ideally, a team will get an alert, track down the issue, remedy it and then drop a line to the FinOps team saying, “We fixed this!” And all the FinOps team has to do is respond with, “Hooray!” The FinOps team goes from being a taskmaster to being a leader paving the way for greater efficiency.
All in all, Ashley showed us how the right FinOps process can truly change the culture of an organization and bring teams together to increase the speed, performance and efficiency of their cloud.
Additional Lessons from the Q&A
With these lessons and FinOps process firmly in mind, the Q&A among members had some great additional insights that built on these ideas:
- Higher Estimates with Lower Goals - Vendors have several pricing tiers, such as with On-Demand and Reserved Instance (RI) pricing. Estimates for budgeting can be set at a higher, On-Demand level, but goals can be set for lower spend that takes advantage of RIs.
- Create Multiple Alert Levels - Some systems let you set multiple alerts with different levels of urgency. Milder alert levels can be used to let teams know when they’re above that lower ideal spend, but still within their overall estimate.
- Consider All Costs in Estimates - The cost of migrating to the cloud goes beyond the compute instance and the various services that go along with it. Things like labor costs, labor savings from managed services and current licensure obligations have to be included.
If you want to hear the whole discussion, along with real-life examples of the FinOps process in action, then join us in the FinOps Foundation!