How do software engineering managers measure their team's performance? Any tools, tips, metrics or suggestions? I'm trying to manage a high performing group & eager to learn from other peoples experience.
Capacity/velocity tied to commitment delivery and their time windows are some of the things I most closely track. Velocity is still considered a top metric for engineering performance. If you are starting out I suggest looking through your tickets or any type of docs/info you can gather on previous work, points, estimates, delivery, etc. to determine where the performance _could_ be. Few thoughts below! Have a lot more but hope this helps.<p>Define an estimate system. We try to keep time out of our estimates when it comes to per story or sub-epics. What's important is getting the team aligned and speaking to the same definitions. Where the variance is the complexity based on the dev's abilities. Juniors you should see a trend of higher estimates typically, seniors more consistent and deliver on it, and the ones who are growing where they can estimate accurately w/ less info or more complex work. (Potential signal they can provide mentoring or might be ready for a promotion.)<p>Capture the data. If you have Jira or something similar its not to hard to generate reports but it means you have to be getting data points like the above down so you can generate things like capacity/velocity over time. Simple spreadsheets are more than fine. You don't have to turn it into rocket science. We initially did a spreadsheet since we couldnt report out of Jira yet until the process caught up with the performance improvements. If no data to start - you take a best guess baseline w/ your team. One of the teams I took on in the past was doing poorly and we didn't have anything. We looked at some previous work/estimates, threw out where we missed deadlines, and said "Ok everyone were going to say we can do 30pts a sprint across the 5 of us. It's slightly lower in general than what we've done but we need a baseline and we'll adjust as we go."<p>- Use the data. You're trying to figure out predictability. It took us about 4 sprints to narrow on what we could complete as a team. We went from about 50-65% hitting our commitments up to 85-95% after adjusting to our new per sprint numbers. Within all these things, the levers and knobs under the hood is looking at actual work more in-depth and how we perceive it and what a good estimate would look like. More accurate estimates stems from growing your business, domain, and technical knowledge. A lot of teams don't actually take the dive into that and just mess around with the number of points. You can't take a passive approach to this as a team and hope it will get better.<p>I just got back from LeadDev eng manager conference and it was HUGE what you're asking about. A few companies there were very focused on this problem space. Swarmia and LinearB were a couple of them. This is going to be a big deal for awhile based on the current economic waves. And it will last long after its over. Great skill to grow in as hard as it can be.