A Better Way to Manage the Cost of Data
Hi, I hope you had a great holiday season! As I promised, let’s talk about how to run a successful data function in companies. Today, I want to start with a very tangible thing almost every data leader deals with, cost.
I really like a quote from an article in “Harvard Business Review” that talks about the role of Chief Data Officer, it says “virtually every problem in a digital economy can be described as a data problem, victory is hard to achieve unless there are agreed-upon metrics against which progress can be measured.” The first metric data leaders should measure, undoubtably, should be the true cost of building each data product. One area we struggle with is managing the operational expense in cloud computing.
A Forrester study conducted in December 2023 found 72% of global companies exceeded their cloud budget in the last fiscal year, despite cloud cost optimization tools & FinOps practices were broadly utilized. The companies in the study responded that their cost optimization toolings and FinOps roles were elusive, reactive, and their remedies couldn’t reach architecture level.
The reason we started building our Data Command Center is that we believe this problem is still not solved. Let me give you an example, there are still organizations goes over their annual cloud budget 6 months into the fiscal year. For them, if you install a monitoring tool that tells their engineering team, 3 months into the year, now you only have half of your budget left, you need to reduce your burn down to 1/3 of your current spend to keep you through the rest of the year, that most likely won’t work. Such a drastric measure could result in significant quality compromise in your analytics products, and serious drifts in your ML models. Maybe a better solution is this, helping them to attribute that cost to each of their product, quantify each product’s impact, share that with the business groups and executives to reinforce a common understanding on data, so next year’s budget can be allocated based on the expected impact of those data products. By following this practice next year, you can align your budget management effort with your engineering effort so it’s easier to control the spend, but also you will shift focus from controlling spend to generating impact instead. In the long run, this changes the dynamic of your data team from an infrastructure team to a value-generation team, that’s how you can really treat data as an asset and a product of business. And that’s one of our approaches to help organizations to resolve their cloud budget overrun problem, it lines our core principle as well.
The reason we can do this is because of our tech,, we are connecting cost and impact for each data product in an automated way by looking at the flow of your data across multiple tools and teams. In the next 1-2 videos, I will highlight how this works in our platform, before I jump into another topic of modern day enterprise data challenges. Okay, until next time, I wish everyone all the best with the rest of your day!
