Using team sentiment and process health metrics to identify developer burnout

In the summer of 2021, we published an article on preventing developer burnout and tech culture’s historical tendency to glamorize being “so busy.” But our interest in the topic is more than a passing trend. Because we know engineering wants to take care of their people, Allstacks has a whole set of features and metrics geared toward keeping a pulse on the team and supporting healthy engineering culture. We analyze unstructured engagement and collaboration data from work patterns, overtime hours (nights and weekends), ticket comments and descriptions, and Slack or Teams to determine the frequency and quantity of engagement as an indicator of team satisfaction and productivity.

Read more about the impact of Covid-19 on software development.

Why track team sentiment?

Attracting and maintaining engineering talent can be difficult. Especially now as demand is growing and companies can offer crazy incentives. So it’s critical to check the pulse of your teams and invest in their personal and professional health while balancing the needs of the business. After all, engineering is a team sport – like a relay – and when one teammate stumbles, it can jeopardize the entire team.

Practically speaking, lower team sentiment can mean a risk of employee turnover, lower code quality, and operational inefficiency – all classic symptoms of employee burnout.

Because burnout has real bottom-line implications, it’s a no-brainer to proactively diagnose and respond to potential risks in your software organization. This is why the team sentiment report exists. It allows team leaders to put early warning signs on something that can otherwise be missed when moving fast, especially with remote work where collaboration can be more difficult.

Allstacks VP of Engineering, Jamie Howard, advocated for this type of insight saying, “Measurement of engineering team health and sentiment is more important than ever with remote. Before, when we were all together in the same space, you could rely on the energy in the room, gut feel, and watercooler talk to gauge how things were going. Remote work means you can’t physically visualize and experience that energy anymore.”

PUTTING DATA BEHIND SENTIMENT

It’s well established that overwork is “literally killing us” and bad for the bottom line to boot, so as a human and a business leader, it makes sense to regularly check your engineering team’s temperature in a concrete and actionable way. Unfortunately, it’s not always obvious how to pin down and address exactly what contributes to burnout in engineering.

As a product or engineering leader, you might gather some anecdotal evidence of your team’s stress around not working on the right things, too many bugs, or tight deadlines – but without data to support your hypothesis, escalating the issue to your business counterparts is likely to fall flat. Putting data behind the sense of unrest can catapult you into solution mode, often before the problem is even noticeable outside your team.

To illustrate, let’s dive into the sentiment report in action with a real example of how our own team recently used Allstacks to address our own challenges around visibility, healthy workload balance, and efficiently delivering strategic objectives – just like our customers.

Recently, we experienced some team growth, introducing new team members, changing up team structure, new processes — all the things. I had a hunch that the team might be fatigued from this and turned to our Team Sentiment report to validate that concern and investigate why.

This was my post in Slack in which we have affectionately coined "the spaghetti monster:"

In which of course the team’s response was...

Joking aside, this helped me validate that my hunch was accurate. I quickly checked in with our backend team to get some insight on the spaghetti monster above, and they candidly shared concerns about bug fixes and triaging overwhelming their workday and pressures to deliver against the strategic initiatives of the business. When I dug deeper and looked into the distribution of work graph, I saw data that corroborated their story, indicated that we had a capacity issue which gave me what I needed to advocate for them.

The Issue Type report reflected that we were nearing 70% unplanned work each month. That not only puts planned efforts and roadmap at risk, but it could put undue pressure on developers to keep up with an unreasonable load of ad hoc tasks.

It was also helpful for me to look at how their work was distributed across projects. I noticed that the team was spending the biggest chunk of time on Uncategorized Tasks, which likely means a lot of the work that was unplanned – putting out fires, emergency fixes, and the like. Because uncategorized work is often technical and tactical rather than a strategic driver, it may put team engagement and delivery of strategic initiatives in jeopardy.

uSING TEAM SENTIMENT & Process Health Metrics TO INFORM DECISION-MAKING

As you may intuit from the story at Allstacks, Team Sentiment and Process Health metrics allow us to be more intentional about what data we look at next in order to spot risk, triage, and address the problem. There are typically three key process health metrics we would turn to for visibility into how we can transform the volatile spaghetti monster into a more stable, predictable current, and happier teams that create value:

1. Workload: This will give you an idea of the balance and intensity of the workload on each team member, which collectively can dramatically impact business outcomes.
2. Epic Focus: This helps you understand what your team is working on (more bugs than features?), and it provides insight into the strategic versus tactical balance of tasks.
3. Work Patterns. This can establish how the team is organizing their lives to get work done. It can show you early mornings, late nights, and weekend work – information that can be used in 1:1 check-ins and to unblock work or inform process improvements.

Team Sentiment is just one dimension, a data-driven way of "feeling the energy of the team" and a better way to evaluate engineering performance while nurturing a healthy team and work culture.

To read more on our point of view around metrics and team building, check out this article.