Every manager who deploys a monitoring tool eventually has the same conversation: what number counts as good? Setting a productivity benchmark is easy. Setting a fair productivity benchmark, the kind that survives a promotion review or a compensation discussion without falling apart, is harder.

Here is the framework we walk new managers through. It is not a magic formula — it is a way to stop the argument before it becomes an HR issue.

Step 1: Pick a unit that matches the work

Time-based benchmarks (hours logged) are the wrong unit for almost any knowledge work. Someone can work seven hours and produce more than a colleague who worked ten. Pick a unit that matches what the team actually delivers:

Engineering: pull requests reviewed, tickets closed, or story points completed
Support: tickets resolved, first-response time
Sales: qualified opportunities created, meetings booked
Content: pieces published, revisions accepted
Design: assets delivered against briefs

The unit does not need to be perfect. It needs to be countable and directional.

Step 2: Establish a baseline before setting a target

Do not set the benchmark first. Measure quietly for four to six weeks, look at the range across the team, and then set the target in the middle third of the distribution — not the top and not the bottom. A target set to your top performer's output demoralizes everyone else. A target set to the bottom is meaningless.

Step 3: Correct for context

Two people doing the same job produce different numbers for reasons that have nothing to do with effort:

One works on a legacy codebase; the other on a greenfield project
One has three days of meetings per week; the other has zero
One handles the on-call rotation this month; the other does not
One is training a new hire; the other is not

Publish the benchmark with the correction rules attached. “This target assumes no more than eight hours of scheduled meetings per week” is a fair caveat. Ignoring the correction and comparing raw numbers is not.

Step 4: Compare to self, not to others

The most useful productivity data is not “you produced less than your peer.” It is “you produced less than you did last quarter.” Individual trajectories are less contentious, easier to explain, and easier to act on. A team dashboard should default to per-person trends, not a leaderboard.

Step 5: Review quarterly, not weekly

Weekly numbers are noisy. Life happens: a sick day, a bad sprint, a customer emergency. Reviewing them week over week produces panic decisions. Quarterly reviews smooth the noise and let the trend show. Use the weekly data for coaching conversations, not for benchmark enforcement.

Step 6: Document the definition

Write down what the benchmark measures, how it is calculated, what the correction rules are, and who signed off. That document becomes the reference in every performance discussion. Undocumented benchmarks turn into arbitrary ones the moment there is disagreement.

What a fair benchmark looks like in practice

A fair benchmark answers “yes” to all of these questions:

Can the employee reproduce the number from their own tools?
Are the correction rules published?
Was the target set from an actual observed range, not pulled from a peer company?
Would you defend the target to the employee's next manager without changing it?

If any of those is a no, the benchmark is not ready to be an accountability tool. It might still be useful as a rough steering signal — but do not use it for compensation or discipline until you can answer yes on all four.

DeskTrust's weekly reports produce the per-person trend view that this framework depends on. See features or start a trial from pricing.

🎯 Setting Fair Productivity Benchmarks: A Manager’s Guide