Time tracking in a small business almost always starts from distrust: it feels like people are working less than they should. It ends either with a screenshot-taking monitor and damaged relationships, or with the whole thing quietly abandoned after a month.
Meanwhile the problem it was meant to solve is usually a different one — understanding what work costs, not catching a slacker. And that is solved differently.
First answer this: why do you want hours?
Everything else follows from the goal. There are four, and they need different levels of precision.
1. Calculating project cost. The most common and most useful. You need hours by project rather than by person, accurate to within half an hour. This pays for itself immediately: without hours, project margin can’t be computed at all.
2. Billing by the hour. Mandatory for hourly contracts. Higher precision, tied to tasks and clients.
3. Planning capacity. How much can we take on next month. A rough estimate is enough — planned hours per person, with actuals secondary.
4. Monitoring people. Here you should stop and think. If that’s the only goal, time tracking won’t achieve it.
Why control through hours doesn’t work
Hours aren’t results. Someone who sat for 8 hours didn’t produce more than someone who closed the same task in 4. Time tracking as a productivity metric rewards slowness.
The numbers adapt. As soon as hours affect how a person is judged, they start reflecting the manager’s expectations rather than the work. And with almost no dishonesty involved: the rounding simply goes the convenient way.
Screenshots and activity monitors destroy trust. They’re the most expensive tool available: they catch isolated cases while telling the entire team “I don’t believe you”. Your strongest people leave first — they have options.
The problem usually isn’t laziness. When someone produces little, the cause is more often unclear tasks, waiting on answers, rework and fragmented attention. Hours will reveal that, but the cure is order in the work, not supervision.
A simple test: if after rolling out tracking you plan to have conversations with people about their hours, you’re building a control system, not an accounting one. Admit that to yourself before you start.
Four ways to count hours
| Method | Precision | Burden on people | Suits |
|---|---|---|---|
| Retrospective estimate (weekly) | low | minimal | broad cost accounting |
| Hours logged per task | medium | small | project work, margin |
| Per-task timer | high | medium | hourly client billing |
| Automatic activity monitoring | falsely high | high + distrust | not recommended |
Retrospective estimate. On Friday everyone splits their week across projects: “40 hours: project A — 20, project B — 15, internal — 5.” Takes three minutes. The error is 10–20%, but that’s enough for cost accounting and there’s almost no resistance.
Hours per task. When closing a task, the person records how long it took. More precise, tied to specific work, and it reveals which tasks systematically cost more than estimated.
A timer. Start before working, stop after. Best precision and worst tolerance for context switching — of which real work has plenty. It makes sense where hours are billed to a client. In our case, for instance, the timer lives inside the messenger mini app: tasks in Telegram and MAX start and stop in two taps, no browser needed.
Automatic activity monitoring. Counts keystrokes, applications, takes screenshots. Formally the most precise; in practice it measures presence rather than work, and costs a great deal in trust. For a team under 50 it’s almost never justified.
Rolling it out without a revolt
1. State the goal honestly and in advance. “I want to understand project costs so we price correctly and stop taking loss-making work.” That’s a clear and respectful reason. If the real goal is supervision, people work it out within a week and the tracking dies.
2. Start with projects, not people. The first question is “how many hours did this project take”, not “how many hours did Igor work”. That difference in phrasing sets the team’s attitude to the whole exercise.
3. Take the least precision that suffices. If a weekly estimate is enough for margin, don’t introduce a timer. Every step up in precision costs resistance.
4. Don’t tie hours to pay right away. The moment hours affect money, they start lying. Let the system run for a month or two as a measuring instrument first.
5. Show the team what came of it. If people see that tracking led to dropping a loss-making client or to revised deadlines in their favour, they’ll maintain it themselves. If the data disappears into a black box, they’ll stop.
6. Agree what counts as work. Stand-ups, learning, client correspondence — those are hours too. If there’s nowhere to log them, people will hide them inside projects and the data will rot.
What to do with the results
After a month you’ll have numbers. The useful conclusions are usually these:
- Real utilisation is lower than you thought. Billable hours are typically 60–75% of calendar hours. That’s normal, not a conversation topic — it just has to be built into your hourly cost.
- Some projects lose money. Usually one or two. Then it’s a decision: raise the price, cut the scope, or part ways.
- Estimates are systematically low. If actuals consistently exceed plan by 30%, that isn’t a people problem — it’s a coefficient to add to your estimates.
- Some time goes to friction. Searching for information, chasing status, rework. What that costs is in our breakdown of the price of task chaos.
None of these conclusions is about someone being lazy. All of them are about how to cost and sell work.
Red flags to avoid
- Requiring people to be online for a set number of hours.
- Screenshots and application monitoring.
- Public leaderboards of hours worked.
- Recalculating pay by hours without warning.
- Demanding minute-level precision.
Each of these buys short-term visibility of control and long-term loss of trust. In a team under 50, trust is worth more.
Bottom line
Time tracking exists to tell you what work costs and how to sell it, not to catch people out. Set the goal honestly, take the least precision that suffices, count by project, show the team the results, and keep hours away from pay for the first few months.
Then it sticks — because it becomes the team’s instrument rather than a supervisor.
In Gosudarynya hours attach to tasks and projects, and margin computes itself: contract minus labour cost. The timer lives in the Telegram and MAX mini app, so no separate service is needed. 7 days free, no card.