How to Track Employee Task Performance

The distinction matters because activity metrics are easy to collect and easy to game, and because they reward the wrong behaviour. Someone who takes on ten tasks and finishes two looks more productive than someone who takes two and finishes both, on any measure that counts events rather than outcomes.

Task outcome data has the opposite property. It is generated as a side effect of doing the work — nobody fills in a form to produce it — and the quantities it measures are the ones a manager is actually accountable for.

Reading them together is what makes them useful. A long turnaround with a short completion time is a queueing problem — the work is fine, it waited. A short time to accept with a long completion time is the opposite: people pick work up promptly and then it takes a while, which may be entirely reasonable.

TaskIt computes, for any scope, how many tasks were assigned, accepted, started and completed, how many are pending, in progress, overdue or blocked, and the completion rate as completed divided by assigned.

TaskIt reports completions and durations for today, the last seven days, the current month and the previous month alongside the overall figures. The comparison between the current and previous month is usually the most informative single number, because it holds the kind of work roughly constant.

Individual figures vary for reasons that have nothing to do with the individual: the mix of work they receive, whether they are the escalation point for a difficult category, whether they were on leave. Department figures average much of that out, and company figures give you the baseline that makes a department number meaningful.

TaskIt computes employee, department and company-wide aggregates from the same task set, and can be filtered by date range, department or individual. How work is attributed to a department is covered in how to manage tasks across multiple departments.

TaskIt produces rankings directly from the computed figures: fastest responder by lowest average time to start, fastest turnaround, most tasks completed, highest completion rate, most completed in the last seven days, and slowest responder by highest average time to start. Each ranking carries the actual value it was ranked on.

• A minimum number of assigned tasks is required to place in the completion-rate ranking, so one task completed out of one does not outrank eighteen out of twenty.
• Someone with no measurable value for a metric — no completed tasks, for example — is simply absent from that ranking rather than counted as zero.
• The slowest-responder ranking exists to surface bottlenecks, not to name and shame. Read it as a question about the queue before it is a question about the person.

• Not time spent. TaskIt has no timer, timesheet or effort log. The elapsed time between two timestamps includes lunch, meetings, other tasks and the night.
• Not difficulty. A four-hour task and a four-hour task are indistinguishable in the data even when one was trivial and the other was not.
• Not quality. Whether the work was any good is a review question, not a timing question.
• Not activity. Nothing here observes what anyone is doing between updates, and nothing should.

Being explicit about this is not a caveat added for form. A manager who believes turnaround measures effort will draw conclusions the data does not support, and the people being measured will notice before the manager does.

1. 01Compare like with like. Compare a person against their own trend, or a department against a comparable one — not two individuals doing different work.
2. 02Prefer trends to snapshots. One slow month is noise. Three is a signal.
3. 03Look at the queue before the person. A rising time to accept usually means too much work arriving, not people ignoring it.
4. 04Ask before concluding. Every unusual figure has a mundane explanation available; find out whether it is the right one.
5. 05Share the figures with the people they describe. Measurement that is only visible upwards is surveillance; measurement everyone can see is feedback.

A task created on Monday, handed over twice and completed on Friday has a four-day turnaround that belongs to the process, not to whoever happened to be holding it at the end. TaskIt counts handovers in and out per person alongside the other figures, which is the context that stops the final assignee being credited or blamed for the whole elapsed time.

That is also why keeping one task ID across a transfer matters for measurement as well as for context — see the task handover process. Reports across tasks, people, workflow, collaboration and recurring schedules are described on the features page.