How to find where your team's hours actually go

The fastest way to find where your team’s hours go is to track work in two buckets for one week: the work only a person can do, and the repeatable work a system could do instead. Most small teams are surprised by how much falls in the second bucket. That gap is where your time is going, and it is also where you get it back.

You do not need a time-tracking tool or a consultant for this. You need one week, an honest list, and a simple rule for sorting what you find. Here is how to run it.

Start with one week, not one day

A single day lies. Mondays look nothing like Fridays, and the busywork that buries a team tends to hide in the edges: the end-of-week report, the month-start reconciliation, the follow-ups that pile up while everyone is heads-down. Track for a full week and the real pattern shows up.

Keep it light. For five working days, have everyone jot what they worked on in rough 30-minute blocks. Not a stopwatch, just a running note. The point is not precision to the minute. The point is to catch the work that never makes it onto a job description but eats the day anyway.

Sort every task into two buckets

At the end of the week, put each task into one of two buckets.

Bucket one is judgment work: the things that need a human. Talking to a customer, making a call on a tradeoff, designing something, deciding what matters this quarter. This is the work you hired your people for.

Bucket two is repeatable work: the things that happen the same way every time. Moving data from one system to another, sorting the inbox, sending the same three follow-ups, formatting the weekly report, copying numbers into a spreadsheet. It has to get done, but it does not need the person doing it.

The line between the buckets is one question: if you wrote down the steps, could someone or something follow them without you in the room? If yes, it is bucket two.

Some tasks will feel like they live in between, and that is fine. A task that is mostly mechanical but needs a person to approve the final step is still mostly bucket two: the system does the work, a person signs off. Do not let a small judgment step at the end keep an hour of repeatable work in bucket one. Split the task at the point where judgment actually enters, and sort each part on its own.

Add up bucket two, honestly

Now total the hours in bucket two. This is the number that matters, and it is usually bigger than anyone expects, because repeatable work is invisible. Nobody brags about the two hours they spent reconciling invoices. It just quietly happens, every week, to your most capable people.

Be honest here, because honesty is the whole value of the exercise. Do not round bucket two down because the tasks feel small. Small and constant is exactly the problem. A task that takes twenty minutes and happens every day is not a small task. It is a part-time job nobody applied for.

Look at who is doing the bucket-two work

The second thing to notice is not just how much repeatable work there is, but who is doing it. On small teams, the answer is almost always your best people, because they are the ones trusted to get it right. That is the expensive part. The person you hired for judgment is spending a chunk of every week on work that needs none.

This is the real cost of busywork, and it does not show up on any invoice. It shows up as the project that keeps slipping, the customer who waited a day too long, the good idea nobody had time to chase. You are not short on talent. You are short on the hours your talent has left after the repeatable work takes its cut.

Why a gut estimate is not enough

You might be tempted to skip the week of notes and just estimate. Do not. The reason the exercise works is that repeatable work is systematically underestimated. Each instance is small, so the brain files it as nothing, and the total never gets added up. Ask someone how much of their week goes to busywork and they will guess low, every time, because they are remembering the big tasks and forgetting the hundred small ones.

Writing it down breaks that pattern. When the same task shows up in the notes on Monday, Tuesday, Wednesday, Thursday, and Friday, it stops looking like nothing and starts looking like what it is. The notes are not there to police anyone. They are there to make the invisible work visible, so you are deciding from a real list instead of a hazy memory.

Pick the three tasks that come back the most

You do not fix all of bucket two at once. Look at the list and find the three tasks that repeat most often and follow the clearest steps. Those are your first candidates to hand off, because they give back the most time for the least effort, and their rules are clear enough that a system can run them reliably.

Good first candidates tend to look the same across teams: inbox triage, moving data between tools, and the routine follow-ups that fall through the cracks. If your week looks like most, at least one of those is near the top of your list.

Decide what to do with the hours you find

Finding the hours is only half the exercise. The other half is deciding what they are for. Before you hand any task off, write down what your team would do with that time back: more hours with customers, the project that keeps getting pushed, leaving at a reasonable hour. That answer is what makes the change stick, and it is the honest measure of whether reclaiming the time was worth it.

This is where an AI agent fits. An agent is well suited to bucket two: the repeatable, rule-based work that runs the same way every time. Hand it the three tasks you found, keep a person in the loop on anything that matters, and your team keeps the judgment work while the agent takes the rest. The goal is not fewer people. It is more of your people’s time on the work that needs them.

Run it again next quarter

Work creeps. New tools, new processes, and new busywork accumulate the same way the old kind did. Run this same week-long check once a quarter and you will catch the repeatable work before it buries the team again. The teams that stay fast are not the ones that never accumulate busywork. They are the ones that keep finding it and handing it off.

The exercise is simple on purpose. One week of honest notes, two buckets, and the three tasks that come back the most. Do that, and you will know exactly where your team’s hours go, and exactly where to get them back.

If you want to see which of those tasks an agent could take on first, Rudder’s AI team plan walks through it in a couple of minutes at wearerudder.com/tools/ai-team.