Activity level is one of the most misread numbers in workforce management. But just because it shows measurable data doesn’t always mean that it equals productivity. And that’s what we’ll debunk in this article.
Most time tracking tools show the number as a percentage next to a person’s name, sometimes colored red or green. Unfortunately, not every manager stops to ask themselves whether that number is just a number or in fact a performance verdict.
So here is what the number is actually made of, what you can reasonably conclude from it, and where it stops being useful.
What activity level actually measures
Activity level measures physical input on a computer. This is keyboard strokes, mouse clicks, and scroll actions that get counted during tracked time and turned into a percentage.
That is the whole definition. The number can’t tell what work was done, what KPIs were achieved, and what went on outside of the device.
What complicates things is that every tool calculates the percentage differently, and the method changes how much you can trust the result. The most common approach checks whether any input happened during each second or each ten-second window, then reports what share of those windows had something in them. Under that method, a person tapping one key every nine seconds can score close to 100%.
How is activity level calculated?
Activity level is calculated by counting keyboard, mouse, and scroll input during tracked time and expressing it as a percentage from 0 to 100. WebWork uses a two-factor formula that combines input intensity, meaning average actions per active minute, with the share of tracked time that contained any input at all.
We built it that way because the single-factor version obviously is flawed. Dividing total actions by total tracked minutes meant that someone who worked intensely for two minutes and did nothing for the following eight could still show a high percentage for that interval. That’s why separating intensity from active time gives a more honest reading of both.
The second thing we changed was the baseline. Activity Sensitivity sets how many actions per active minute count as full input intensity, and you can choose between Relaxed at 25 actions per minute, Normal at 50, and Strict at 100. Relaxed suits designers, managers, and researchers. Strict suits data entry and high-frequency computer roles.
Without that adjustment, every role in your company gets graded against a typing speed that only applies to some of them.
What activity level can tell you
There are three things that activity level does honestly.
- It confirms that tracked time was working time. For example, you often need evidence that the hours on the timesheet involved someone at a keyboard.
- It shows deviation from a person’s own baseline. When someone who normally runs at 55% drops to 20% for a week, that is worth a conversation. The conversation is usually about blockers, workload, or wellbeing, not about effort.
- It shows patterns at team and workspace level. For example, if activity across a whole department collapses every Thursday afternoon, the cause is far more likely to be a recurring meeting block than fifteen people losing motivation at the same time.
All three of these use activity level as a signal about presence and engagement and not actual productivity or output.
If you want to see how activity data looks when it is calibrated by role rather than applied flat, you can try WebWork free for 14 days.
What activity level can’t tell you
Now, let’s take a look at the limits, because this is where most reporting goes wrong.
- Thinking and reading barely register or don’t register at all
Reviewing a pull request, reading a technical spec, working through a contract, sitting on a client call produces very little input and all of it is real work.
- Roles produce completely different baselines
A support agent might handle twelve chats while an architect might draw one system diagram. That is exactly why you can’t use the same baseline for different professionals.
- AI-assisted work has changed the arithmetic
A writer who prompts a model and then edits the output produces a fraction of the keystrokes they would have produced writing from scratch (I’m telling you this from experience). But they will still ship the same or even more work. The same shift is happening in code, research, and reporting across most teams right now. Input volume is falling for reasons that have nothing to do with effort but rather external reasons.
Can employees fake activity levels?
Yes. Mouse jigglers and auto-clickers simulate input, and simple percentage-based activity metrics cannot distinguish simulated movement from real work. This is the most common criticism of input tracking, and it is a fair one.
What catches it is pattern analysis rather than the percentage itself. WebWork’s Unusual Activity Tracking flags three specific signatures: activity above 95% sustained for 45 minutes or longer, input fluctuation below 4% across an hour, and heavy mouse usage paired with almost no keyboard input.
The feature is available on the Premium plan.
Note that when people start faking activity, the problem is not the faked number but the root cause, which is most likely the unhealthy practice of tracking. So that’s something worth observing and preventing.
What is a good activity level percentage?
There is no universal good percentage. A realistic target depends on the role, the work being done that week, and the sensitivity setting behind the calculation.
WebWork’s default ranges put anything below 30% in the Low band, 30% to 60% in Medium, and above 60% in High. Those thresholds are a starting point, and you can adjust them in settings.
One more thing to consider is the trend line instead of the absolute percentage.
Why is my activity level low?
Low activity usually means the work involved reading, thinking, watching, or talking rather than typing and clicking. Meetings, video calls, phone calls, and long documents all produce low input. Remember that sensitivity settings that are too strict for the role will also push the number down so it’s worth discussing with the workspace manager.
If the reading looks wrong, check the sensitivity setting before you check the person.
How to use activity level without damaging trust
Four rules cover most of it.
- Never use activity level on its own in a performance review, a pay decision, or a termination. It is one reading among several, and it is the least connected to results.
- Set sensitivity by role before you set any expectation.
- Give employees access to their own data. When people can see the same numbers their manager sees, the metric becomes something they can work with.
- Read it next to output such as tasks closed, projects delivered, tickets resolved, revenue booked. Activity level describes the input side of the day, and output metrics describe what the day produced.
There is a compliance argument for the same approach. Several jurisdictions have moved toward requiring employers to disclose specifically which monitoring methods they use, rather than issuing a blanket notice that computer activity may be monitored. If you have to name what you collect and explain why, it becomes much harder to justify a metric you are using outside its actual purpose.
The metrics worth reading next to activity level
Work Mode Overview separates Focus Time, meaning stretches of concentrated work without distraction, from Shallow Time, meaning lower-intensity work. Two people at the same activity percentage can have very different focus profiles.
Work categorization classifies apps and websites as Productive, Neutral, or Non-Productive. WebWork AI handles this automatically after ten minutes of use, and the classifications are team-aware. For example, LinkedIn counts as productive for a marketing team and neutral for a development team, and the tool has to be able to tell those apart.
Productivity Insights brings the time breakdown, productivity score, app and website usage, engagement metrics, and office versus remote comparison into one view, alongside activity level against the workspace average.
Task and project data closes the loop. Hours tracked against a project, tasks completed, budget consumed against estimate.
Read together, these tell you how a week actually went. Read alone, an activity percentage tells you how much someone touched their keyboard.
Where it’s going
As more of the mechanical work moves to AI tools, input volume becomes a weaker and weaker signal about productive work. Instead, it should be used in unison with outcomes and produced work as a measure of engagement.
We are building in that direction too, which is why our AI features focus on interpreting patterns rather than counting clicks harder. Activity level stays in the product, and it stays useful, as long as everyone reading it knows exactly what it covers.
If you want to see how this works with your own team’s data, WebWork is free for 14 days and you do not need a card to start.
Try Activity Level Tracking on WebWork for Free
