Measure the wrong thing and a team will optimize it. That is the entire risk of productivity metrics, and it explains why most dashboards make teams worse.
- ›Why activity metrics backfire
- ›Five metrics worth tracking
- ›The metrics to stop reporting
- ›Reading them without damaging trust
Why activity metrics backfire
Tasks completed, messages sent, hours logged, and commits made all measure motion. They are easy to collect, easy to compare, and easy to game, usually without anyone intending to game them.
A team measured on tasks completed will split work into smaller tasks. A team measured on hours will log hours. Neither is dishonesty; it is the predictable response to being measured on a proxy.
Five metrics worth tracking
Cycle time. How long work takes from start to done. It captures real flow, it is hard to game without actually going faster, and it surfaces the queues where work waits.
Work in progress. How many things are in flight at once. High WIP predicts long cycle times reliably, and reducing it is one of the few changes that works quickly.
Time in review. The most commonly missed queue in any team. Work sitting in review is finished effort producing zero value.
Carryover or spillover rate. What share of committed work does not finish. Consistently high carryover means commitment volume is wrong, which is a fixable planning problem.
Unplanned work share. What proportion of the period went to things nobody planned. This one names the real cause of missed commitments in most teams, and it usually surprises people.
What to stop reporting
- ›Individual task counts. They compare incomparable work and invite gaming.
- ›Hours logged, unless you bill by the hour. Then it is a billing input, not a productivity metric.
- ›Velocity as a target. Velocity is a planning input. The moment it becomes a goal it inflates, and it inflates by estimate rather than by output.
- ›Utilization alone. Push it toward 100% and you eliminate the slack that absorbs variability, which makes everything slower. Pair it with cycle time or margin.
Important: Report team-level, not individual-level, unless you are managing a specific performance issue with a specific person. Individual dashboards published to the team change behavior in ways that outlast whatever they were meant to fix.
Reading them without damaging trust
Metrics are for finding questions, not for answering them. Cycle time doubled last month is a question, and the answer might be a hard problem, a departure, or a metric artifact.
Share the numbers with the team before sharing them upward. A team that sees its own data first treats metrics as information; a team that learns about them in a leadership meeting treats them as surveillance.
AI and metrics
Where AI helps is explanation rather than collection. Any tool can chart cycle time. Reading across projects to say which changes coincided with the shift, and which work types drive it, is where a cross-module AI adds something a chart does not. See AI business intelligence.
FAQ
What is a good cycle time?
There is no universal number. Your trend is the signal; comparisons to other teams are noise.
How often should we review these?
Monthly for trends, and at retrospectives for context. Weekly review produces reaction to normal variation.
What if leadership wants individual metrics?
Offer team-level metrics plus the specific evidence when there is a genuine individual concern. Publishing individual productivity dashboards is one of the reliably damaging management practices.
Brainis tracks cycle time, work in progress, and unplanned work across Work OS, with Cortex explaining the changes. See pricing.
Sharing insights on business operations, AI, and modern team management.
Run your company on Brainis
All 11 Operating Systems on every plan, from $29 a month. No per-seat pricing — you pay for AI capacity, not headcount.
See pricing