How to tell if remote employees are actually working

Are they actually working?

When your team is remote, the honest answer to “are they actually working?” comes from two places: the work that exists, and the signals that explain it. You can’t walk past a desk and see concentration, or notice the Friday afternoon drift. What you have instead is the work that arrives, and, if you use monitoring software, a set of signals about how the days were spent.

Start with the concession that matters most: the work itself is the final test. Someone who ships on schedule doesn’t need an alibi, and no activity chart should put them on trial. Monitoring earns its place when work slows down and you need to see why — and even then, the data explains, it doesn’t judge. This post walks the signals you’ll see, why any one of them misleads on its own, and how to read them together, including the case every manager asks about first: the mouse jiggler.

What each signal means

Activity level is the bluntest signal. It counts mouse movement and keyboard use, and answers one narrow question: is someone at the computer? It cannot tell you whether what happened at the computer was work. Filling in a spreadsheet and scrolling a feed keep the number up all the same.

App and website usage gets closer to the answer. It shows where the hours went — the editor, the ticket board, the documentation site, or social media. A day’s usage often lets you tell a sales day from a build day at a glance. But six hours in the right app still isn’t output, and some real work — a long phone call, a whiteboard, a printed draft — never touches an app or website at all.

Idle time is the stretch where nothing moved. It can be lunch, a school run, or a hard problem that needed away-from-keyboard thought. Offline work leaves no trace in activity data, so a run of idle time means “ask”, not “caught”.

Screenshots add the context the numbers lack. An activity number says two hours went somewhere; screenshots show what was on the screen while it went. They also keep the other signals honest, because a record of real work looks different, hour after hour, from a record that repeats.

The daily timeline puts the pieces in order: the day broken down to productive time, idle time and break time, hour by hour. One signal is a fragment. A day laid out in sequence starts to tell a story.

Why one signal alone misleads

Take any single number and it will mislead you in both directions.

The mouse jiggler is the case managers ask about first. It’s a small device or script that moves the cursor so the computer never looks idle. In activity data, a jiggler looks like engagement: the activity level stays high all day. But it fakes one signal, not all of them. Hours of cursor movement with no keyboard use, the same window on screen for half a day, screenshots that repeat or show nothing new, and no delivered work — a day like that doesn’t add up, and the disagreement between the signals is the tell.

Be precise about what that is and isn’t. It’s a pattern you can check for in data like this, not a verdict: the signals show you a day that doesn’t add up, and the conversation that follows is yours to have. It also isn’t always about theft — plenty of jigglers mask breaks rather than steal hours, which is a management problem, not a forensic one. And someone determined to fake more than cursor movement can, so treat the data as a way to narrow a question, never as the answer to it.

The opposite error is just as common: reading low activity as not working. Calls, client visits, reading, planning — the work that moves a project forward often leaves no keyboard trace at all. A quiet chart can belong to your most productive person. This is why no single signal settles anything: high activity can hide a jiggler, low activity can hide your best week.

The work settles it

Before you read any signal, name the output you expect: what should exist at the end of a week — tickets closed, drafts written, releases shipped, customers answered. That list is the test the signals serve. When the output is there, the signals don’t matter; let people work. When it’s missing day after day, the signals turn “I think something’s off” into a specific, fair conversation: here is the output gap, and here is where the time actually went.

The usual objection to monitoring is “track results, not activity”, and it is right as far as it goes. Results come first. But in some roles results go quiet for weeks, and by the time a miss is visible in the output, the weeks that explain it are gone. Activity data is how you see the trend before it becomes a quarter.

Monitor with the team’s knowledge

However you read the data, read it in the open. Choose visible mode over stealth, tell the team what is collected and why, and keep the purpose on proof of work rather than gotchas. A team that knows the tool exists and what it shows will treat the numbers as a shared record; a team that discovers them later will treat every question as an accusation. Workfolio runs in visible mode or stealth mode — choose visible.

Reading a day as a whole

This is the read a tool like user activity monitoring software exists for. Workfolio tracks app and website usage, lays out each day on a timeline broken down to productive time, idle time and break time, takes recurring screenshots up to every 1 minute, and can auto detect slacking employees with rules you customize — the signals above, in one place, attached to the people they describe.

Then run the sequence in that order: define the output you expect, read the signals against each other, and act only where the two disagree. For most of a good team, the question “are they actually working?” answers itself in the work. The signals are for finding the days, and the people, where it doesn’t.