Metrics that support judgement
A count is not automatically a measure of value. More tickets closed, messages sent or hours recorded can coexist with poor quality, invisible rework and unhappy users. Build a small measure set that explains whether the intended outcome improved.
Begin with the purpose and a baseline. Choose a few measures for outcome, quality, timeliness, cost or effort and user or staff experience. Segment where context differs, pair numbers with feedback and set a review date. Do not use a single activity count as proof of improvement.
Build a balanced scorecard
Outcome
Completion, resolved need, correct decision or service delivered as intended.
Quality and safety
Error, rework, complaint, escalation, exception or control failure.
Time and flow
Lead time, wait time, predictability and age of unfinished work.
Effort and cost
Appropriate active effort, cost and capacity rather than raw busyness.
Experience
Feedback from the people receiving and doing the work, including accessibility.
Unintended effect
Work displaced, new workarounds, privacy risk, overload or quality loss.
Write a measure card before collecting data
- Question: which decision will this measure inform?
- Definition: what exactly is included, excluded and counted?
- Source and owner: where does it come from and who checks its quality?
- Baseline and segment: what was true before the change, and which groups differ?
- Review rule: what result would cause the team to keep, change, pause or stop?
Use numbers with lived evidence
The GOV.UK Service Standard says performance data should be combined with user research and used to decide what to improve. Its guidance is for public services, but the principle helps any organisation avoid a dashboard that cannot explain the outcome.
Productivity measurement FAQs
What is a vanity metric?
It is a number that looks positive but does not reliably show whether an important outcome improved. Volume, clicks, messages or hours can be useful context, but they need quality, consequence and user evidence.
Should targets be shared with staff?
Share the purpose, definitions, limitations and how data will be used. If a target creates fear or can be met by harming quality, people will naturally work around it rather than improve the system.
How often should productivity be measured?
Use a cadence that fits the work. Fast-changing operations may need frequent operational signals; slower work may need a longer period to reveal meaningful quality and outcome trends.
Continue your productivity decision
Use the next guide that fits the bottleneck, measurement, work design, communication, attention, automation, knowledge or review question still in front of you.