One of the key conversations organizations and teams forget about in their transition to agile and DevOps is updating their measurement and metrics plan. Many companies are still using measurements and metrics from the traditional waterfall software development lifecycle. While some of these remain useful, many may not provide value to the team or organization—and some may even adversely impact progress toward goals.
A culture of continuous improvement demands that we constantly reevaluate all dimensions of our project, processes, and products, identify what’s working, prioritize opportunities for improvement, and incrementally improve. Our measurements and metrics are no exception; these too must be constantly evaluated to ensure that the information on our dashboards is providing teams the timely feedback they need to track product quality, status, effectiveness, efficiency, resources, and issues.
When transforming to agile and DevOps, we must also transform our measurement and metrics approach. Step back for a moment and qualitatively ask yourself, on a scale of 1 to 5 (5 being excellent), how beneficial your current measurements and metrics are in these areas:
If there are areas where you believe your current metrics fall short, here is one approach to rethinking your measurement and metrics plan.
First, start by understanding the business goals at the company, program, product, and project levels.
Next, link your current measurements and metrics to these overarching goals. You will most likely find some gaps, overlaps, and misalignments—that is, areas where your existing measurements do not support (or even conflict with) the stated business goals, or areas where information is being reported on dashboards that have no relationship to the overarching goals. This mapping exercise will help you better understand what changes you need to make to your measurement and metrics system.
As you begin to reformulate your measurement plan, I recommend using the GQM+Strategies approach. The idea is that we must first define the goals, then develop the questions that need to be answered in support of the goals, and from the questions we can then create the measurements and metrics. This methodology increases the probability that we have the right measures in place, and just enough.
Alternatively, you can map your current measurements and metrics from the bottom up. Take one of your metrics, specify what questions the metric is attempting to provide answers for, and then link those questions to the goal. Whether you work this top-down or bottom-up against your current measurement and metrics system, you will most likely find opportunities for improvement.
Measurements and metrics do indeed drive behavior, so having the right “dials” in place is critical to successfully creating an agile and DevOps culture and delivering quality value to our customers.
