Engineering productivity, metrics, & morale

You will find below my talk (sponsored by the non-profit https://big-on.dev/ in Singapore) on engineering productivity, metrics, and morale.  The questions were awesome. 😎

It's all about psychological safety for the team for the best combination of team morale and team effectiveness.

Summary

Development metrics

• Team productivity: Per department and per team merge request (AKA pull request) rate, defined as the number of requests merged divided by the number of employees on that team
• Productivity of the review process: Open merge request review time
• Quality: Past due issues by severity and type (customer-facing, security, infrastructure)
• Web application performance: LCP (largest contentful paint) 
• Backend application health: Error budgets (error rates and performance analysis per service/endpoint)
• Investment: Issue count by weight by type (feature, bug, maintenance, other)

GitLab-specific development metrics

Team Morale

• Use a combination of sync (meeting) and async (in writing) discussions, with async being the priority.
• Turning on video during meetings is highly encouraged.
• Recording meetings when possible
• Always explain the "why" and not just the "what"
• Release retrospectives (what went well, what didn't, what actions to take)
• Quarterly social activities
• Social slack channels
• Skip level meetings and AMA (ask me anything) sessions
• NPS (net promoter score) for employees

Remote and results

• Measure results, not hours
• Hire where the best people are (vs in central locations)
• Written processes 
• Written knowledge
• Transparent sharing to increase collaboration
• Anyone can contribute

Video

Slides