How do I know whether my engineering team is performing well?
Short answer
Look at whether engineering helps customers and the business achieve their goals, delivers changes predictably and keeps the product dependable. Combine outcome measures with delivery and operational evidence, then discuss the constraints with the team. Commit counts, hours worked and a busy roadmap cannot establish performance on their own.
Start with the outcome, not a league table
Choose a customer journey that matters: completing a purchase, joining an event or producing a useful result. Establish whether it works, whether customers use it, and what it costs to support. A team can ship frequently while improving none of these things.
Compare a team with its own previous performance and changing responsibilities. Comparing unrelated teams without accounting for their products, dependencies and risk encourages the wrong competition.
Use a small set of questions
- Are the changes we deliver improving an agreed customer or business outcome?
- How long does a small change take, and where does it wait?
- How often do changes cause problems, and how quickly can we recover?
- How much time goes into incidents, manual work and avoidable rework?
- Can the team explain its priorities and the tradeoffs behind them?
Delivery measures help locate constraints; they do not replace judgement about value. Pair each number with an explanation and a decision it could change.
DORA: software delivery performance metrics (opens in new window)
Use the evidence to support the team
If delivery slows, examine changing priorities, external dependencies and unclear decisions before attributing it to individual effort. Agree one improvement, its owner and a signal that would show whether it helped.
At Remo I led engineering through growth, contraction and acquisition. Those changing conditions are why I connect engineering performance with the business situation rather than treating more output as the objective.
Related questions
Should we measure individual developer productivity?
Use individual conversations to understand contribution and support needs. Team delivery statistics are not a reliable ranking of individuals, particularly when people spend time mentoring, reviewing or preventing incidents.
When is outside help useful?
When leaders cannot agree what good performance means, or the same constraints persist despite attempts to address them. Start by understanding the situation and whether an adviser, coach or direct leader would help.