Your Time Tracking Data Shows Where Your Margin Goes
Almost every agency owner I talk to has months, sometimes years, of time tracking data sitting in Harvest or Clockify that nobody’s done anything with.
The team logs their hours, the invoices go out, and the data sits there. It got set up for billing, or for keeping an eye on utilization, and once it did that job nobody went back to it.
That’s a shame, because it’s the best source you have for figuring out where your margin is going.
Your P&L can’t see what your time tracker can
Your P&L tells you what you made and what you spent. It can’t tell you which account cost you money this quarter, or which service is eating hours nobody planned for. Everything gets blended together into one margin number, and a blended number hides the accounts and services pulling it down.
Your time data is already broken out the way you need it. Every entry has a person, a client, a project, and usually a task. Line that up against what each client pays you and what each service is supposed to take, and the margin problem stops being a feeling.
When I break it out by service and by account with the agencies I work with, it shows exactly which changes need to be made. I’ve seen the crown jewel account turn out to be the one running at the lowest effective hourly rate. I’ve seen a service that looked totally fine on the P&L lose a big chunk of its hours to third and fourth rounds of revisions nobody was counting.
Neither of those showed up anywhere else. They were sitting in the time data the whole time.
Why nobody looks
Until recently, getting to that view meant a BI tool, a consultant, or a very patient ops person with a big spreadsheet. Most agencies at $3-5M don’t have the first, don’t want to pay for the second, and their ops person is already buried.
So the export never happens, and the questions that matter most to the owner get answered by gut feel. Which client should we raise at renewal? Which service needs to be rebuilt? Where are the senior hours going? Everyone has a hunch. Nobody has the number.
Now you can do it yourself in an afternoon with Claude.
What it takes
You need three files from your own systems. The first is a time export, every entry with the person, the client, the task, and the hours. The second is your services menu, what you sell and how many hours each deliverable is supposed to take. The third is your monthly fees by client.
From there it’s three prompts. One matches every time entry to a service on your menu. One rolls the hours up by service and by account and compares them to what was sold. One builds the dashboard, with hours against budget by service, the revision hours past what the menu includes, and each account’s effective bill rate.
I tested every one of those prompts on a sample agency’s 5,000-plus time entries before I trusted them, because the first version of any prompt like this gets things wrong in ways that look right.
The things that will throw your numbers off
A few things in the data will skew the answer if you don’t catch them. Two cause most of the trouble.
The first is prepaid clients. If a client pays for a quarter up front, the money lands in one month and the hours spread across three. Look at that account month by month and it swings from wildly profitable to underwater and back. You have to spread the fee across the months it covers before the rate means anything.
The second is time off nobody logs. If vacation and sick days never make it into the tracker, it looks like people have hours they don’t have. Your capacity numbers come out wrong, and every utilization number built on them goes with it.
Neither of those is hard to fix once you know to look for them. They’re the reason a dashboard can look finished and still be wrong.
“Our time tracking isn’t good enough for this”
I hear this one constantly, and it’s usually not true. The data doesn’t have to be perfect to show you the pattern. If one account is running at half the rate of the others, a few missing entries won’t hide that. If revisions are eating a third of a service’s hours, you’ll see it even when half the team logs loosely.
What messy data does is make the first pass imprecise. That’s fine. An imprecise view of where the margin goes is a lot more useful than no view at all, and it tells you exactly where your tracking needs to get better.
The bigger risk is waiting for clean data before you look. That day doesn’t come. Meanwhile the crown jewel account keeps running at the lowest rate, and the revision rounds keep piling up.
What to do with it once you have it
The dashboard isn’t the point. It shows you where to go next. Pick the one service or account where the numbers look worst, sit down with the people who deliver it, and find out why the hours run over. Usually it’s a step nobody owns, a brief that sends drafts the wrong way, or a senior person doing something a junior could do.
Fix that, run the numbers again next quarter, and you’ll see it move. You already paid for the data when your team logged every hour of it.