Skip to content
Engineering Lead Handbook

Measuring success

The metrics that show up in dashboards are usually the ones that are easy to count. That’s not the same as the ones that matter. These are the pulse checks I use — team, individual, org — that don’t fit in a dashboard but tell me whether the engineering org is actually healthy.

Team pulse

  • Is velocity trending up, down, or flat over the last quarter? Flat with the same people is a signal — either the team plateaued, the process ossified, or the work got harder in a way nobody named.
  • If a new hire cloned the repo today, could they get it running from the README in under an hour? If not, it’s rotted. Fixing the README is a two-hour job someone should own this week, not “eventually.”
  • Are standups still short? If they’ve stretched to 30 minutes, they’ve turned into status meetings and stopped being for the team.
  • Are retros blameless and real? The failure modes: retros that stay polite avoid the actual problem, and retros that get blunt turn into blame. Watch for both; you need the honesty without the finger-pointing.
  • When did we last delete code? Teams that only add code are teams that don’t understand their own system.

Individual pulse (from a manager’s seat)

  • Do they know what the next level looks like? If they can’t tell me what they’d have to do to get promoted, I haven’t done my job.
  • Does feedback land both ways? Am I hearing from them what I’m doing wrong, or just delivering opinions in one direction?
  • Are they doing work only they can do? If yes, that’s a growth risk (bus factor) and a delegation opportunity. If no, they’re not being stretched.
  • When did I last publicly credit them for something? If I have to think, it’s been too long.

Org pulse

  • Who’s ready for promotion in the next six months? If I can’t name them, I’m going to lose someone to a company that could.
  • Who’s stuck? What is one thing I could do this quarter to unstick them? Not “I’ll think about their growth” — one specific thing.
  • Where’s the single point of failure? Whose leaving would take out a system? That’s the next thing to invest in redundancy for.
  • What are we NOT doing that we should be? The absence of a decision is a decision. Named absences are easier to defend than surprise gaps.