Summary
Every productivity figure in monitoring software rests on a table that says which applications and websites count as productive, which are neutral, and which are unproductive. Most organisations set that table once, company-wide, in the first week, and never look at it again. The result is a design team whose core tool sits in the neutral column, a developer whose documentation site is marked unproductive, and productivity percentages that are wrong by design for everyone whose job differs from the person who set them up. This guide gives a method that produces accurate classifications: categories per role rather than per company, a starter list that covers most knowledge work, explicit rules for the tools that are productive for one person and a distraction for another, and a review loop that keeps the table honest as tools change.
E-Monitor supports per-role categories and lets employees flag a misclassified application from their own dashboard. See application tracking →
What the Categories Do
Monitoring software records the application in the foreground and the website in the browser, minute by minute. On its own that is a long list of names. The classification table turns the list into the three numbers most dashboards lead with: productive time, neutral time and unproductive time, and from those a productivity percentage. Everything downstream, from the team report to any productivity score, inherits the table's accuracy.
The three categories are blunt by design. Productive means the tool is where the role's work gets done. Unproductive means the tool has no plausible work use for the role. Neutral is everything else: tools that are necessary but not the work itself, such as email and chat, and tools whose use is sometimes work and sometimes not. A fourth category, uncategorised, is where new tools land until someone decides, and it should be close to empty.
The mistake is to treat neutral as a mild form of unproductive. It is not. Email is neutral because reading it is not output, but a day with no email would be a day of blocked colleagues. The limits of the data guide explains why the category of a tool says nothing about the quality of what happened in it.
Why Company-Wide Classification Fails
One table for the whole company encodes one job's view of work. The person who set it up, usually in IT or operations, marked their own tools productive, left the rest neutral, and marked social media unproductive. For every other role the table is wrong somewhere.
| Tool | Company-wide default | Social media manager | Developer | Recruiter |
|---|---|---|---|---|
| Unproductive | Productive | Neutral | Productive | |
| YouTube | Unproductive | Productive | Neutral (tutorials) | Unproductive |
| Stack Overflow | Neutral | Unproductive | Productive | Unproductive |
| Figma | Neutral | Productive | Neutral | Unproductive |
| Slack | Neutral | Neutral | Neutral | Productive (candidate chat) |
| Spotify | Unproductive | Neutral | Neutral | Neutral |
With the default table, the social media manager's productivity figure is wrong by half and the developer's by a third, and the team comparisons built on those figures are meaningless. Per-role tables fix this and are supported by most tools, including E-Monitor, through group or profile settings.
The Starter List
A starting point for knowledge-work roles, to be adjusted per team. The right column is the one that needs most attention, because it is where most errors live.
| Category | Productive | Neutral | Unproductive |
|---|---|---|---|
| Creation tools | Editors, IDEs, design tools, CAD, spreadsheets, document editors, CRM records | ||
| Communication | Customer-facing: ticketing, support chat, dialler | Internal: email, chat, video meetings | |
| Reference | Documentation, internal wiki, code hosting | General search, encyclopaedias | |
| Planning | Project and task tools | Calendar | |
| Admin | HR portal, expenses, timesheets, password manager | ||
| Learning | Role-specific training platforms | Video sites when used for tutorials | |
| Media and social | Only for roles that work in them | Music and podcasts | Social feeds, video entertainment |
| Personal | Banking, health, government services | Shopping, games, gambling, adult content |
Two deliberate choices in the list. Music and podcast services are neutral, not unproductive, because they run in the background during work and a rule that marks them unproductive produces hours of false unproductive time. Personal admin such as banking is neutral because adults handle their lives during the day and a rule that penalises it is a rule people resent and route around.
Rules for Ambiguous Tools
Some tools cannot be classified by name because the same site is work for one minute and not the next. Four rules handle most of them.
Classify by role, then by section where the tool allows it. YouTube can be split by channel or search term in some tools; LinkedIn by feed versus messaging. Where that split exists, use it; where it does not, choose the category that matches the majority use for the role and accept the error.
When in doubt, neutral. An uncertain tool marked neutral does no harm to anyone's figures. The same tool marked unproductive produces a wrong figure that someone has to argue about later.
Never mark a tool unproductive for a role that has asked for it. If a team says a site is part of the job, the burden is on the organisation to show otherwise, not on the team to prove it.
Uncategorised is a queue, not a category. Review it weekly for the first month and monthly after. Anything sitting in it for more than a month is a tool the organisation is using without having decided what it is for.
Involving the Teams
The fastest route to an accurate table is to ask the people who use the tools. Give each team lead the draft classification for their role before launch and a week to correct it. Then, after launch, give every employee a way to flag a misclassified application from their own dashboard. Most tools support this; E-Monitor routes the flag to the administrator with the employee's note attached.
This is not only about accuracy. A classification employees helped set is a classification they accept, and the conversation about which tools count as work is one of the more useful conversations a team can have about how it works. The employee's guide to the dashboard tells them to check the categories in the first week; the organisation should expect the flags and welcome them.
Keeping It Accurate
Tools change. A new AI assistant appears in half the team's day within a quarter; a project moves from one platform to another; a role changes shape. A classification table that was right in January is wrong by June unless someone owns it.
- Owner: one named person, usually in operations or IT, with team leads as reviewers
- Uncategorised review: weekly for a month after launch, then monthly
- Role review: quarterly, with each team lead confirming their table
- Change log: every reclassification dated, because it changes historical comparisons
- Reclassification note on dashboards: when a tool moves category, the trend line should say so, or the move will read as a change in behaviour
The change log matters more than it sounds. Moving a tool from neutral to productive raises the productivity figure for everyone who uses it, and a manager who does not know the table changed will read the jump as a real improvement. The trends guide covers how to read a line that has a reclassification in it.
A Worked Example: The Design Team's Missing Hours
A twelve-person product design team showed a productivity figure of 54 percent, lowest in the company, three months after monitoring launched. The team lead's instinct was to challenge the software. The administrator's instinct was to challenge the team. The classification table settled it.
The company-wide table had been built by the operations manager. Figma, the team's primary tool, was uncategorised and therefore excluded. The design-system documentation site was neutral. Dribbble and Behance, used daily for reference, were unproductive. Loom, used to record design walkthroughs for engineers, was unproductive because it was a video site. Reclassified per role, the same three months of data gave the team a figure of 79 percent, second in the company, with no change in anything the team had done.
The lesson was general. Every team's table was reviewed with its lead that month, four other roles had similar gaps, and the review became quarterly. The false positives guide covers the other common ways monitoring data is wrong before anyone has done anything.
Frequently Asked Questions
1. How do monitoring tools decide what is productive?
Through a classification table that assigns each application and website to a category: productive, neutral or unproductive. The software records which application is in the foreground, and the table turns that into productive time and a productivity percentage. The figures are only as accurate as the table.
2. Should app categories be set per company or per role?
Per role. A company-wide table encodes one job's view of work and is wrong for every other role: LinkedIn is unproductive for a developer and productive for a recruiter. Most tools support per-role or per-team tables through group settings.
3. What should be classified as neutral?
Tools that are necessary but are not the work itself, such as email, chat, calendar and HR portals, plus anything whose use is sometimes work and sometimes not, plus music and podcast services and personal admin such as banking. When in doubt, neutral.
4. How do you classify YouTube or LinkedIn?
By role first, and by section where the tool allows it, for example channel or feed versus messaging. Where no split exists, choose the category that matches the majority use for that role and document the choice. Never mark a tool unproductive for a role that has said it is part of the job.
5. Can employees correct a wrong classification?
They should be able to, from their own dashboard, with the flag routed to the administrator. Employees are the fastest route to an accurate table, and a classification they helped set is one they accept.
6. How often should app classifications be reviewed?
The uncategorised queue weekly for the first month and monthly after; each role's table quarterly with its team lead. Every reclassification should be dated in a change log, because it alters historical comparisons.
Categories that match the job E-Monitor supports per-role application categories, routes employee flags to the administrator, and keeps the uncategorised queue visible, so productivity figures describe the work actually done. Sign up →