This blog is the first in a two-part series and was adapted from The Enterprisers Project.
In 2020, a year like no other, is it still useful to measure IT value based on green, yellow, or red lights on a screen?
Now that infrastructure is everything – powering productivity, cutting OPEX, and supporting digital initiatives that may change overnight – flashing lights on a monitor are no longer enough to keep the wheels moving.
It’s time to develop new metrics based on an increasingly digital world."
While availability is still profoundly important – count on users to complain loudly when an app or site is down or glitching out – traditional IT metrics such as server capacity, I/O, utilization, and network throughput are now table stakes for survival.
In organizations with hefty cloud-based investments, those data center metrics are less relevant because infrastructure components are abstracted and delivered as a service.
Cloud infrastructure is measured on response time, scalability, security, and cost per customer/user."
Another trend supporting new metrics is the advent of intelligent, automated infrastructure monitoring systems that abdicate the need to supervise those flashing lights all day.
Forrester Research analyst Rich Lane writes about the problem with old metrics in his blog: “Measures such as MTTR are antiquated in environments where systems have been built to be highly resilient and automatically scalable. If I&O is doing its job correctly, the low-complexity, high-volume incidents are being rooted out of the system.”
From servers to sales and satisfaction
Do you have the right metrics to demonstrate value for users and stakeholders? Do you know how (and why) users prefer one service or app over another? Can the work of IT help frontline employees make better decisions to grow revenues, loyalty, and customer satisfaction?
I’d like to say yes, yes, and yes. But this takes some work and reorientation.
Thinking about the new metrics
Start by uncovering the critical use cases for driving organizational success. These are the digital initiatives and technologies that support customer-facing products and services along with the tools that enable employees to best serve your market. With this knowledge, IT operations can begin creating the new metrics and correlating them to the “old metrics” of availability and performance.
Old and New ITOps metrics
| Old | New |
| CPU utilization | Items per order |
| IOPS | Time to complete an online transaction |
| Capacity | Service/subscriptions per customer |
| Latency | Digital interactions per customer |
| Network throughput | Cross-sell/upsell revenue |
| MTTD/MTTR | Employee satisfaction rates with digital tools |
| Time to acknowledge/respond | Employee productivity (invoices approved, contracts created, programs completed, reports generated, customer issues resolved, etc) |
Let’s look at some examples:
E-commerce: For online retailers, the focus of the last several years has been understanding buyer intent: What might they buy and when? Which factors play into the customer leaving the site and not buying or returning as a repeat customer?
Therefore, metrics could include the number of items purchased in an order, the dollar value of an order, shopping cart abandon rates, and transaction time from when the customer hits the site until purchase.
Understanding these metrics could inform decisions such as: Do web pages load fast enough? Is our recommendations engine working properly? Are products out of stock or is the purchasing process difficult? Do we need to optimize cloud services or select new ones to support all of the above?
Banking: Cross-selling and upselling is a vital revenue-building tactic in financial services. Therefore, one might want to measure the number of banking products and services consumed per customer.
Another measure of customer satisfaction is the quality of digital services. How many electronic transactions and interactions are customers conducting per week and how does that correlate to lower attrition or increased revenue per customer? Again, negative results could indicate IT performance issues while conversely, positive results confirm the current strategy and technology portfolio.
Automotive: In-vehicle technologies like OnStar can send alerts to customers based on their driving behavior and let them know when their oil needs to be changed or when it’s time for recommended maintenance. This is an example of customer value in the form of issue avoidance and user experience. Measuring the accuracy and impact of these monitoring technologies would be valuable, especially if connected to customer satisfaction metrics and total cost of ownership for the vehicle.
In part 2, I’ll share how to begin implementing these new metrics in your IT ops organization.
Next Steps:
- Read about the 5 Tips for Observability Success
- How to Define Awesome SLOs with OpsRamp
- Predictive Analytics in IT Operations
