Summary
A productivity baseline is the measured normal for a team or a person: what the numbers look like in an ordinary week, with the variation that ordinary weeks contain. Without one, every figure on a dashboard is compared against a guess, and the guess is usually the manager's own working pattern or a number from a vendor's benchmark page. With one, a change is visible as a change, and the size of the change can be judged against the size of normal variation rather than against zero. This guide gives the method: how long the baseline window should be, which weeks to leave out, how to build a team baseline and an individual one, how to express the normal range so that noise is not read as signal, and when to throw the baseline away and start again.
E-Monitor keeps the trailing weeks on every dashboard, so the baseline and the current week sit side by side without a spreadsheet. Sign up →
Why Every Number Needs a Baseline
A manager opens a dashboard and sees that an analyst's focus time last week was four hours. Is that good? The honest answer is that nobody can know without the analyst's usual figure, the team's usual figure, and the range both usually move within. Four hours is a collapse for someone who normally does twelve and a strong week for someone who normally does two, and both are common.
Published benchmarks do not solve this. The benchmarks page is useful for sizing a team against its industry, but a benchmark describes a population, and a person is not a population. The only comparison that supports a decision about a specific team or person is that team's or person's own history. That history, cleaned and summarised, is the baseline.
Most of the harm done with monitoring data traces back to the missing baseline. Single-day call-outs, rankings against the top performer, and surprise findings in reviews are all comparisons against the wrong reference. The trends versus snapshots guide covers the reading side; this guide covers building the reference first.
Step 1: Choose the Window
Two to four ordinary working weeks is the right length. Two weeks is the minimum for a weekly pattern to appear twice; four weeks smooths a single odd week without reaching back into a different season. Longer windows feel more rigorous and are usually worse, because they average across changes in headcount, tooling and workload that make the result describe nobody.
The window should be recent, continuous, and chosen before the data is looked at. Picking the window after seeing the numbers is how a baseline becomes an argument. For a new monitoring deployment, the first fortnight is often distorted by the deployment itself, as people notice the agent and adjust; start the baseline window in week three. The implementation guide builds this delay into its rollout plan.
Step 2: Exclude the Weeks That Are Not Normal
The baseline is supposed to describe an ordinary week, so weeks that are not ordinary are removed before averaging. The usual exclusions:
- Public holidays and weeks with a holiday in them, which compress four days of work into a different shape
- Quarter-end, month-end and year-end for finance, sales and any team with a closing cycle
- Launches, migrations and incidents that pulled the team off its normal work
- Weeks with significant leave, where the team's totals are low for reasons that have nothing to do with the team
- The first two weeks after a reorganisation, a new manager or a tooling change
If exclusions leave fewer than two usable weeks, extend the window backward rather than accept a thin baseline. Record what was excluded and why, because the exclusion list is the first thing a sceptical employee or works council will ask about, and a documented list answers the question.
Step 3: Build the Team Baseline
For each metric that matters, take the weekly values across the usable weeks and record three things: the median, the typical range, and the weekday shape.
The median rather than the mean, because weekly figures have outliers and the median ignores them. The typical range as the band between the lowest and highest usable week, or, for teams with more than six weeks of history, the band one standard deviation either side of the mean. The weekday shape because most teams have a predictable Monday and Friday, and a Friday that looks like a Wednesday is a signal in its own right.
| Metric | Median week | Typical range | What falls outside it means |
|---|---|---|---|
| Active hours per person | 34.5 | 31 to 37 | Below: leave, blockers, disengagement. Above: overload |
| Focus blocks per person | 11 | 8 to 14 | Below: meeting creep. Above: usually a calendar change |
| Meeting share | 24% | 19% to 29% | Above: process or manager change |
| After-hours activity per person | 1.2 h | 0 to 2.5 h | Above: workload set for more hours than exist |
| Output (tickets, deliverables) | 62 | 52 to 71 | Read against active hours, never alone |
The metrics are the ones from the productivity metrics guide; the values are illustrative. The range column is the one managers skip and the one that prevents most mistakes: a week at 32 active hours against a median of 34.5 is inside normal, and nothing needs saying about it.
Step 4: Build Individual Baselines Separately
An individual's baseline is built the same way from their own weeks, and it is theirs. It is not compared with the team median to produce a ranking, and it is not compared with another person's. Its purpose is to make a change in the person's own pattern visible, which is the only comparison that is fair to someone whose role, chronotype and calendar differ from everyone else's.
Show each employee their own baseline on their dashboard. People who can see their normal range stop worrying about a single low day, and they are the first to notice when something in their week has changed. The employee's guide to the dashboard explains what they will see. The individual versus team guide covers who should see which baseline.
Step 5: Decide What Counts as a Change
A single week outside the typical range is noted, not acted on. Two consecutive weeks outside the range in the same direction is a pattern worth a question. Three is a change, and the question becomes what changed in the work.
Write the rule down before the data arrives, because it is the rule that stops a manager reacting to a bad Tuesday. The rule applies symmetrically: three weeks above the range on active hours is as much a signal as three below, and usually the more urgent one. The first-time manager guide turns the rule into a fifteen-minute weekly routine.
When to Reset the Baseline
A baseline describes the conditions it was measured under, and it expires when those change. Reset it after a reorganisation, a change of manager, a tooling or process change that alters the application mix, a significant change in headcount, or a move between office and remote patterns. Reset it on a schedule too, quarterly at most, because a baseline that is a year old is measuring a team that no longer exists.
Resetting is not the same as moving the goalposts. The old baseline is kept, and the difference between old and new is itself a finding: if the new normal has fewer focus blocks than the old one, the reorganisation cost the team something, and that is worth knowing.
A Worked Example: Reading One Week Two Ways
A seven-person content team had a week in which active hours fell to 30 per person, focus blocks fell to 7, and output fell to 48 pieces. The manager's first reading was that the team had a bad week and something needed saying.
Against the baseline, the reading changed. The team's typical range for active hours was 31 to 37, so 30 was just outside normal. Focus blocks at 7 were below the range of 8 to 14. Output at 48 was below the range of 52 to 71. Three metrics outside range in one week was worth a question. The question found that a client review cycle had landed on the Tuesday and Wednesday, putting most of the team in review meetings for two days. The following week every metric was back inside its range.
Without the baseline the manager would have had a conversation about effort with a team that had spent two days doing what the client asked. With it, the only action was to note the review cycle as a future exclusion. The limits of the data guide covers why that distinction matters.
Frequently Asked Questions
1. What is a productivity baseline?
The measured normal for a team or a person: the median weekly value and the typical range for each metric, taken from two to four ordinary working weeks. It is the reference every later figure is compared against, so that a change is judged against normal variation rather than against zero or against someone else.
2. How long should a productivity baseline period be?
Two to four ordinary working weeks, recent and continuous, chosen before looking at the data. Longer windows average across changes in staffing and tooling and end up describing nobody. For a new monitoring deployment, start the window in week three to avoid the adjustment period.
3. Which weeks should be excluded from a baseline?
Weeks with public holidays, quarter-end or month-end cycles, launches, migrations, incidents, significant leave, and the first two weeks after a reorganisation or tooling change. Document the exclusions.
4. Should baselines be per person or per team?
Both, built separately and used differently. The team baseline supports scheduling and workload decisions. The individual baseline makes a change in that person's own pattern visible and should be shown to them. Neither is used to rank people against each other.
5. How much change is significant?
One week outside the typical range is noted. Two consecutive weeks in the same direction is a question. Three is a change, and the question becomes what changed in the work. The rule applies equally to figures above the range and below it.
6. When should a productivity baseline be reset?
After a reorganisation, a change of manager, a tooling or process change, a significant headcount change, or a shift between office and remote work, and in any case quarterly. Keep the old baseline, because the difference between old and new is a finding in itself.
Every number next to its normal E-Monitor shows the trailing weeks and typical range beside each metric, for the team and for each employee, so the baseline is always on the screen. Sign up →