Many small teams know how much a job brought in, but not how much of their own time it took. A quote is based on a feeling from the last similar job. A job that "went fine" turns out, months later, to have taken twice the planned hours. Nobody can say which jobs were worth doing, because nobody wrote the hours down.
Tracking time does not need special software or a long form. It needs a small agreement about what to record, when to record it and what to do with the result. This article describes a minimal version for a team of two to twenty people.
Decide why you are tracking
The method depends on the purpose, so name it before you start. The most useful reasons for a small team are:
- Pricing. You want to know how long typical jobs really take, so that the next quote is based on facts.
- Planning. You want to see who is loaded and who has room, before you promise a date.
- Billing. You charge by the hour and need a record to show the customer.
Pick one main reason. If it is pricing, you can be relaxed about fifteen minutes here and there. If it is billing by the hour, the record has to be more careful, and the customer may ask to see it. Tracking time to check up on individual people is a different purpose, and it tends to spoil the numbers: people start to round their entries to look busy or to hide slow days.
What to record
Three facts are enough for each entry:
- The job. Use a short name that everyone has agreed, such as the customer name and a word for the work ("Novak – website update"). Keep a single list of current jobs in one shared place. If people are free to invent names, the same job will appear under several spellings and the totals will be wrong.
- The person. Who did the work.
- The time. How long it took, in blocks of fifteen or thirty minutes.
Add a fourth column only if you will actually use it. A common useful one is a short type of work, for example "making", "meetings with the customer" or "corrections". It lets you see later how much of a job went on corrections, which is often the most surprising number. A free-text note on every entry is rarely read and slows people down.
Also keep a place for internal time that is not billed to any customer, such as admin, training and sales conversations. Without it, people squeeze that time into customer jobs, and the jobs look more expensive than they were.
How often to record
Write time down the same day. A few minutes at the end of the day, or at the end of each piece of work, is enough. Entries made from memory a week later are mostly guesses, and they are consistently too low for the work that was interrupted and too high for the work that was memorable.
Do not try to track every five minutes. Fifteen or thirty minutes is precise enough for decisions about pricing and planning. A day that is mostly one job can be one entry.
A simple shared spreadsheet is enough to start with: one row per entry, with columns for date, person, job and time. Once you have more than a few hundred rows, or want totals per job without effort, a simple time-tracking tool may be worth it. Choose it by how quickly a person can add an entry, not by how many reports it offers. If adding an entry takes more than half a minute, people will stop.
Make it a habit, not a campaign
Nobody enjoys recording time. Three things help:
- Explain the purpose. Say what the numbers will be used for and what they will not be used for.
- Go first. If the owner or manager records their own time too, the practice is seen as normal rather than as supervision.
- Remind lightly. A short message on Friday afternoon about missing days is enough. Do not chase people daily.
How to use the numbers
The record only pays off when you look at it. Once a month, take thirty minutes and answer four questions:
- Which finished jobs took longer than estimated, and by how much? Put the estimate and the actual hours side by side. A difference of ten percent is normal. A difference of fifty percent deserves a closer look.
- Why? Look at the type of work. If corrections or customer changes explain most of the difference, the problem may be a vague brief, not slow work. If meetings explain it, the next quote should include them. For ideas on what a fixed-price estimate should contain, see a fixed-price estimate for a small software job.
- Which kinds of job are worth doing? Divide what you charged by the hours spent. Jobs that bring noticeably less per hour than the rest are candidates for a higher price, a narrower scope or a polite refusal.
- Is the load balanced? If one person has a full month and another has room, change who gets the next job before you promise a date.
Write the answer to each question in one line, and change one thing for the next month: a figure in your price list, a standard allowance for meetings, or a rule about how many rounds of corrections a quote includes. Without a change, the exercise is only reporting.
If you want to see what unrecorded time can cost in an hourly business, professional services losing billable hours gives a closer look. For a quick way to put a value on repeated manual work, see the cost of one manual process.
Common problems
- Entries are missing. Look at the days with nothing recorded before asking why. It is usually a forgotten habit, not a refusal. A shorter form often fixes it.
- Everything is recorded as "general". The job list is too long or too unclear. Cut it to the jobs that are actually running.
- Numbers look too good. If every job comes in exactly on estimate, people may be adjusting their entries to match. Say again that the numbers are for pricing, and that a job taking longer than planned is useful information rather than a failure.
- Time for holidays and sickness. Keep it out of the job totals, but note it somewhere, because it affects how many hours a person can actually offer next month.
A short start
In the first week, agree the purpose and the job list, and create the shared sheet. In the first month, only record. Do not draw conclusions from fewer than three or four finished jobs. At the end of the first month, hold the first thirty-minute review and change one figure. After three months you will have a record that makes quoting noticeably less of a guess.