Every tool in this category calls itself simple, which makes the word useless for choosing between them. It becomes useful again if you define it as something countable, and there is an obvious candidate: the number of decisions one entry costs you.
A simple time tracker is one where recording a switch takes a single decision and a single action. That matters because the switch happens during the work, not after it, and a method that asks four questions at that moment is a method you will stop using by the second busy week. Everything else called simple is really just uncluttered.
The distinction is worth drawing carefully, because simplicity is a trade rather than a virtue, and what it gives up is specific.
How do you count the decisions in an entry?
Walk through what recording one switch actually asks of you, method by method.
The counts below are this article’s own tally rather than a measured figure, so treat them as a way of comparing methods rather than as data. A paper log asks three: what to call this, when it started, and, later, when it ended. Plus the physical act of writing, which is fast but requires the notebook to be within reach.
A spreadsheet asks four or five: find the file, find the right row, type a start, type an end, type a label. The row-finding is the part people underestimate, because it happens while you are still thinking about the thing you just stopped doing.
A full-featured tracker usually asks five or six: pick the client, pick the project, pick the task, write a description, set the billable flag, then start. Every one of those fields exists for a real reason, and every one of them is a question asked at the worst possible moment.
A one-tap tracker asks one: which activity is this. Starting it and stopping the previous one are the same gesture.
Now put a number on the difference. Take a modest six switches a day, which is 6 x 5 = 30 switches in a working week. At one decision each that is 30 decisions a week. At five decisions each it is 30 x 5 = 150. Across a 47-week working year that is 30 x 47 = 1,410 against 150 x 47 = 7,050, a difference of 5,640 small interruptions to your own train of thought.
The seconds matter less than that count, but they are worth running too. If a five-field entry takes 45 seconds and a one-tap entry takes 3, the week costs 30 x 45 = 1,350 seconds, or 22.5 minutes, against 30 x 3 = 90 seconds, or 1.5 minutes. Over 47 weeks that is 1,057.5 minutes, about 17 hours 38 minutes, against 70.5 minutes, about 1 hour 11. Those figures assume the entries are made at all, and that assumption is the whole problem.
Why does the moment of entry decide everything?
Because the alternative is reconstruction, and reconstruction is a different and worse operation.
Observational research on information work found the average time in a coherent unit of work was 11 minutes 4 seconds before a switch or interruption, that 57 percent of those units were interrupted, and that returning to interrupted work took an average of 25 minutes 26 seconds when it happened the same day. A day shaped like that is not one where you will remember the sequence on Friday.
Bodies that depend on accurate records say the same thing from the other direction. The State Bar of California’s advisory on bill padding notes that contemporaneous records are almost universally acknowledged as best practice, and warns that “the longer the time period between when the task was performed and the corresponding time entry is recorded, the greater the probability that the time entry may lack accuracy”, adding that a reconstructed record can look extremely precise while “the appearance of accuracy may be deceptive”. That is US guidance for lawyers’ fee bills rather than a rule that binds you, and the mechanism it describes is general.
So the test of a tracking method is not what it can produce at the end of a month. It is whether an entry survives being made at 11:07 on a Tuesday when something has just gone wrong.
What does simplicity actually lose?
A lot, and pretending otherwise is how these comparisons become useless.
It loses structure. No client, project and task hierarchy means you cannot ask “how long did phase two of the Henderson job take” unless you made an activity for it in advance. A flat list of activities answers questions you thought of before the week, not after it.
It loses money. No rates means no revenue figure and no invoice, and those have to happen wherever they already happen.
It loses other people. A single-user tracker has nothing to say about a team, and personal time tracking and team time tracking are genuinely different jobs with different requirements. A tool that does one well is usually bad at the other.
It loses retrospective detail. A one-line activity name is all you will have. If you need entry-level descriptions of the sort the bar advisory describes, a tracker built around a description field is the right tool and the extra decisions are the price of it.
Those are real losses. The reason to accept them is that a detailed record you did not keep is worth less than a coarse one you did.
Paper, spreadsheet, full-featured or one-tap
Judged on attributes rather than a score, because a score would just be this author’s weighting dressed up as a fact.
| Paper log | Spreadsheet | Full-featured tracker | One-tap tracker | |
|---|---|---|---|---|
| Decisions per entry | 3 | 4 to 5 | 5 to 6 | 1 |
| Best at | working with no device at all | restructuring the data afterward | client, project and rate detail in one place | surviving a fragmented day |
| Loses | arithmetic, and the page is often elsewhere | the moment of entry, because it is filled in later | the moment of entry, because it asks too much at it | structure you did not set up in advance |
| Needs an account | no | no | usually | depends on the app |
| Output | pages to add up | a sortable grid | reports and often an invoice | totals by activity, read on screen |
Every column wins something genuine. Paper needs no battery. A spreadsheet is the only one where you can invent a new question in March and answer it against January’s data. A full-featured tracker is the right answer when the detail is the point, and the extra fields are not clutter, they are the product.
How to choose based on how your own week goes
Count your switches for two days. Not your tasks, your switches, including the ones that last ten minutes.
Under about three a day, almost anything works, and a timesheet template or a notebook will do the job without further thought. Between three and six, a spreadsheet holds up if you fill it in the same day, with the arithmetic caveats in time tracking in a spreadsheet. Above six, the method has to cost about one decision or it will not survive the fourth busy week in a row.
Then check what you need out of the other end. If the record has to become an invoice inside the same tool, you want the detailed one and should accept its entry cost. If the record has to become a total you can trust, you want the cheap one. Freelance time tracking covers the routine either way.
None of the four is ruled out by the rules that do exist. The US Department of Labor’s recordkeeping fact sheet states that the Fair Labor Standards Act “requires no particular form for the records” and that any timekeeping plan is acceptable as long as it is complete and accurate, and the US and UK tax authorities are equally indifferent to the medium of a self-employed person’s own records. Time tracking in a spreadsheet quotes all three. Format is not the constraint. Completeness is.
Timeness is the one-decision column, and nothing beyond it. Custom activities, one tap to start or switch, a day timeline, and day, week, month and year donut charts. A note on each entry, data on the device with private iCloud sync, no account, and no export, so the totals are read on screen rather than opened in a spreadsheet. It has no clients list, no rates, no invoices, no projects hierarchy and no team features, because adding them would add the fields this whole page is about avoiding. The money stays where you already invoice, and the arithmetic stays in the hours calculator on this site.
If your week is three long blocks, do not install anything. If it is twelve switches and a Friday spent reconstructing them, the decision count is the only specification that matters.
Sources
- US Department of Labor Fact Sheet 21: Recordkeeping Requirements Under the Fair Labor Standards Act
- Internal Revenue Service: Recordkeeping for small businesses and the self-employed
- GOV.UK: Keeping your pay and tax records
- State Bar of California Arbitration Advisory 2016-02: Analysis of Potential Bill Padding and Other Billing Issues
- Mark, Gonzalez and Harris (2005), No Task Left Behind? Examining the Nature of Fragmented Work, CHI 2005, University of California, Irvine