It is Monday morning and the weekly timesheet report is open in front of you. One designer logged 31 hours instead of 40. A developer put 14 hours on a task estimated at four. Someone logged nothing on Thursday afternoon. You can feel the questions forming, and you can also feel how they will land if you ask them the wrong way.

This is a common bind for team leads and agency owners. You need the numbers to run the business: to bill clients, protect margins and spot projects going off track. But the moment the timesheet feels like a register of who was at their desk, people start logging time to look good rather than to be accurate. Then the report stops being useful to anyone.

The good news is that you can get almost everything you need from a timesheet without inspecting anyone's day. It comes down to what you look for, and what you do with it.

Why timesheet reviews drift into surveillance

Most managers do not set out to micromanage. It usually happens for three ordinary reasons.

The report is organised by person, so you read it by person. A timesheet grouped by team member invites you to judge each row. Who is low? Who is high? That framing turns a project question into a performance question before you have thought about it.

Hours are the easiest thing to measure. Outcomes are harder to see at a glance, so attention slides towards totals. But a full 40 hours tells you very little about whether the right work got done.

There is no agreed purpose for the review. If the team does not know why you read the timesheet, they will assume the least generous reason. Silence from a manager after a timesheet review often feels more worrying than a direct question.

None of this is solved by reading the report less carefully. It is solved by reading it for a different purpose.

Decide what the report is for before you open it

Write down, for yourself and ideally for the team, the two or three questions the weekly timesheet is meant to answer. For a small agency or project team, they are usually some version of:

  • Is the time going to the projects and clients we expected this week?
  • Is any task or project taking far more effort than we planned?
  • Is anyone overloaded, or consistently pulled onto unplanned work?
  • Is the billable time ready to invoice, and does it match what we agreed with the client?

Notice what is missing: "Did everyone work enough hours?" If attendance genuinely matters in your business, handle it as a separate, explicit policy. Mixing it into project review is what makes the timesheet feel like surveillance.

Once the purpose is clear, tell the team. A short message such as "I review the timesheet weekly to catch tasks that are over budget and people who are overloaded, so we can fix the plan" changes how every later question is received.

Read for patterns, not for minutes

With the questions set, change how you slice the data. In Tilvin, the timesheet report can be filtered by date, project and person, and the reports area also includes billable hours and team workload views. Use those filters to start from the project, not the individual.

Start with projects, then tasks

Filter the week by project first. Compare the hours on each project with what you expected. A retainer client that normally takes 15 hours a week but absorbed 32 is worth understanding. Then drill into the tasks behind that total. Because time logs sit against tasks, and tasks carry estimates, you can see which specific piece of work grew rather than guessing which person was slow.

Look for the shape of a problem

A single odd day rarely means anything. Patterns do. A few worth watching:

  • Repeated overruns on the same type of task. If every "homepage revisions" task runs at three times its estimate, your estimates or your revision process need work, not your team.
  • Time spread thinly across many projects. Someone logging 45 minutes on eight different projects in a day is probably being interrupted constantly. That is a scheduling issue.
  • Hours logged at project level with no task. A growing pile of time logged against the project rather than a task often signals unplanned work that never made it onto a task list.
  • A steady gap between billable and total hours. If non-billable time keeps creeping up on a client, there may be scope that was never priced. The routine in Billable vs Non-Billable Hours: A Time Logging Routine for Busy Weeks helps keep those flags honest.

Check workload before you check individuals

Only after the project view do you look at people, and even then through the team workload report rather than raw totals. The question is not "who logged the least?" but "who is carrying too much, and who has room?" Low hours often mean someone was blocked, waiting on a client or covering work that never got logged. High hours across several weeks usually mean a planning problem that will become a retention problem.

Turn what you find into a conversation, not a correction

The way you follow up decides whether the timesheet builds trust or erodes it. A few habits help.

Ask about the work, not the hours. "The API task ran well past its estimate. What did we not see coming?" invites an explanation. "Why did you spend 14 hours on this?" invites a defence. The data is the same; the outcome is very different.

Ask in the place where the work lives. If a task overran, put your question in that task's discussion and @notify the person. The context stays with the task, others who touch it later can see what happened, and it reads as a question about the work rather than a private summons. If the thread gets long, an AI summary of the discussion can pull out the open questions so you can decide what to change.

Fix the plan, visibly. When a pattern turns out to be a bad estimate or unplanned client work, act on it. Adjust estimates on similar tasks, add the missing work to the task list, or talk to the client about scope. When people see that timesheet questions lead to better plans rather than blame, they log more accurately.

Leave small gaps alone. A missing hour on a Thursday is almost never worth a message. Save your attention for patterns that affect delivery, margin or wellbeing.

If you want to take this further, Read Your Project Health Report Before the Client Does covers how to pair time data with project health so you spot trouble earlier.

Next steps: a 20-minute weekly routine

You do not need a long review. A consistent, short one works better. Try this every Monday for a month:

  1. Filter the timesheet report to last week and group your reading by project. Note any project noticeably above or below expectation.
  2. For each outlier, open the tasks behind it. Compare logged time with estimates and look for repeated overruns or time logged without a task.
  3. Check the team workload report. Flag anyone who looks overloaded for a second week running, and anyone who seems underused or blocked.
  4. Check billable hours for clients you invoice soon. Confirm the billable flags look right before the numbers reach an invoice.
  5. Write no more than three follow-ups. Put each one on the relevant task as a question about the work, and note one planning change you will make.

After a few weeks, share what you have learned with the team: which estimates you corrected, which client conversations happened, which workloads you rebalanced. That closes the loop and shows the timesheet is a planning tool, not a scorecard.

The report itself is neutral. Read it for patterns, ask about the work, and change the plan when the data tells you to, and your team will start to see it the same way.