Time tracking has a reputation problem, and it is earned. Most people's experience of it is a fortnight of diligent timers followed by three weeks of blank days, ending in an afternoon spent reconstructing a month from memory and calendar entries. Then the conclusion: this does not work for me.
What actually failed was not the discipline. It was the design. Hours were recorded against nothing in particular, so they answered no question, so nobody looked at them — and something nobody looks at is impossible to keep doing.
Decide what an hour attaches to first
This is the decision that determines whether tracking is useful, and it gets made by accident more often than on purpose. There are three levels, and the right one depends on what you intend to do with the numbers.
| Hours attach to | Answers | Cost |
|---|---|---|
| The client | Am I making money on this account | Almost none |
| The project | Did this job come in over or under | Low, and enough for most people |
| The task | Which parts of the work I consistently underestimate | Real, and only worth it if you quote fixed prices |
Most freelancers should track to the project and stop there. Task-level tracking is genuinely valuable if you sell fixed prices, because it is the only thing that improves an estimate — but it is also the level at which people give up, because it asks you to categorise every twenty minutes of your day.
A running timer, or writing the day up afterwards
Two methods, and the honest comparison is not the one that tool vendors make.
- A running timer is accurate and fragile. It records what actually happened, and it breaks the first time you forget to start it, take a call, or work somewhere without the app open.
- Writing up the day at the end is robust and approximate. You will never miss a day, and you will round everything to the nearest half hour in your own favour without noticing.
In practice the combination wins: a timer when you are working in blocks on one thing, and a five-minute write-up at the end of any day the timer did not survive. What loses, every time, is reconstructing a week on Friday. A week reconstructed on Friday is a week of fiction, and its worst property is that it looks like data.
Worked hours and billable hours are different columns
This is the distinction that turns tracking from an accounting chore into something that changes decisions. Worked is what you did. Billable is what the client agreed to pay for. They are never the same, and the gap between them is where your effective rate goes.
Record both against every entry, and decide the billable flag when you record it rather than at invoicing time. Deciding at invoicing time means deciding under time pressure, in bulk, on entries you no longer remember — which is exactly when unbillable work gets quietly billed and billable work gets quietly written off.
- Rework caused by your own mistake: worked, not billable. Deciding this at the moment it happens is much easier than three weeks later.
- Rework caused by a changed brief: billable, and flag it as a change while the client still remembers asking.
- The meeting where the change was requested: billable if you bill meetings, and you should have said so in the quote.
- Learning a tool the job required: worked. Whether it is billable is a question you answer once, in your terms, not per entry.
Divide last year's invoiced total by your recorded worked hours and you get your effective rate, which is the only rate that describes your business. It is usually well below your headline rate, and the gap is made of exactly the entries above.
Write the entry for the person who will question it
Every entry has two readers: you, next month, working out where a project went; and a client's finance department, deciding whether to approve a line. Both of them need the same thing, which is a sentence that means something without you in the room.
- "Development, 6h" tells nobody anything and invites a query on the total.
- "Checkout: fixed VAT rounding on discounted items, 6h" ends the conversation before it starts.
- Name the deliverable, not the activity. "Meeting" is an activity; "Scope call for phase 2" is a deliverable.
The second version takes four seconds longer to write and saves the twenty minutes you would spend reconstructing it under pressure. It also changes what the client argues about: a specific line invites a question about the work, a vague one invites a question about the number.
Turning hours into something a client approves
The final step is the one most setups skip, and it is where tracking either earns its keep or becomes an internal ritual. Hours have to leave your tool in a form the client can look at and say yes to, before they become an invoice.
Do that and two things change. Disputes move from after the invoice to before it, which is a completely different conversation — you are asking someone to confirm a summary, not to justify a charge. And approval becomes a fact with a date on it, so the invoice that follows carries an answer to "what is this for" already attached.
What to look for in a tracker
The buying guide covers the questions this post assumes: what an hour can attach to, who approves it, whether unbillable time is first-class, and how the hours get from the timer into an invoice.