A time audit is a short, structured log of what you actually spend your working hours on, kept for one typical week and then totalled by task. It is the fastest way to find out which repetitive jobs are worth automating, because it replaces a guess with a ranked list. This guide gives you a log template you can copy, a summary sheet, a worked example with invented numbers, and a method for turning the results into an automation shortlist. The research quoted comes from published studies and surveys, and each is labelled with its source and limits.
Why log instead of guess
Most owners can name the tasks that annoy them. Fewer can say how many hours a week those tasks take, and the annoying ones are not always the expensive ones. A log shows which recurring jobs add up.
Published research gives a sense of the scale, but not your numbers.
- The McKinsey Global Institute's 2012 report on social technologies estimated that the average high-skill knowledge worker spends about 28 percent of the workweek managing email and nearly 20 percent looking for internal information or tracking down colleagues who can help.
- Microsoft's 2023 Work Trend Index, a survey of 31,000 people in 31 countries, found that email, meetings and chat took up an average of 57 percent of the workday, leaving 43 percent for creative work. The figures are self-reported.
- Gloria Mark's research, reported by Gallup, shadowed 36 managers, analysts, developers and engineers for three days. Interrupted work that was resumed the same day was picked up again after 23 minutes and 15 seconds on average. That is a small, specialised sample, so treat it as an illustration of the cost of switching tasks, not a rule for your team.
These averages come from other populations and are several years old. They suggest that coordination work is large. They cannot tell you where your hours go, which is what the audit is for.
What you need
- One typical week. Pick five working days without holidays or a launch deadline.
- A place to log. A spreadsheet or a notes file. Keep it open all day.
- Two minutes per entry. Log when you switch tasks, or at the end of each hour, not at the end of the day. Memory at 6 pm is unreliable.
- A fixed set of categories, so that entries can be totalled.
Suggested categories:
| Category | What it covers |
|---|---|
| Delivery | The work clients pay for |
| Sales and marketing | Enquiries, calls, proposals, content |
| Admin and coordination | Scheduling, data entry, status updates, invoicing, file searching, chasing |
| Management | Planning, hiring, supervising, finance review |
| Other | Learning, breaks that ran long, anything that does not fit |
The log template
Copy this table into a spreadsheet. One row per task or per block of time. The last two columns are what make the audit useful for automation.
| Date | Task (short, verb first) | Category | Minutes | Recurring? (Y/N) | Rule-based? (Y/N/Part) |
|---|---|---|---|---|---|
| Mon | Add new enquiry to CRM and send first reply | Sales and marketing | 8 | Y | Y |
| Mon | Write weekly status update for Client A | Admin and coordination | 25 | Y | Part |
| Mon | Search inbox for last quote sent to Client B | Admin and coordination | 4 | Y | N |
| Tue | Rebuild proposal outline | Sales and marketing | 40 | N | N |
Two rules keep the log honest:
- Write the task, not the project. "Update Client A status" is useful. "Client A work" is not.
- Mark a task Rule-based only if you could write down the rule. If it depends on gut feel, mark it N. If part of it follows rules, mark it Part.
The summary sheet
After five days, build two summaries.
By category. Total the minutes per category. This shows the shape of the week: how much goes to delivery, how much to admin.
By recurring task. Group the entries that are really the same task, and total them. Then add three columns: times per week, minutes each and hours per week. Sort by hours, largest first.
A worked example
The numbers below are invented to show the method. Imagine a five-person service agency where one person logs a week of admin.
| Recurring task | Times a week | Minutes each | Minutes a week | Rule-based? |
|---|---|---|---|---|
| Log new enquiries and send first reply | 10 | 8 | 80 | Yes |
| Book and reschedule calls | 12 | 6 | 72 | Yes |
| Weekly client status updates | 6 | 25 | 150 | Part |
| Chase unpaid invoices | 4 | 10 | 40 | Part |
| Set up folders and forms for a new client | 2 | 45 | 90 | Yes |
| Search for past files and emails | 15 | 4 | 60 | No |
| Copy data between tools | 10 | 5 | 50 | Yes |
| Total | 542 |
The total is 542 minutes, or about 9 hours a week. Four tasks are clearly rule-based: enquiries, booking, client set-up and copying data. Together they take 80 + 72 + 90 + 50 = 292 minutes, which is about 4.9 hours a week, or roughly 21 hours a month. The status updates and invoice chasing are partly rule-based, so they are candidates for a draft-and-approve automation rather than a fully automatic one. Searching for files is a filing problem, and the fix is a naming convention and a shared folder structure, not automation.
Two cautions apply to any result like this. First, minutes saved are only worth money if you spend them on something valuable. Second, an automation needs build time and upkeep, so the net saving is always smaller than the gross one.
Turn the results into a shortlist
- Take the top recurring tasks by hours. Ignore one-off tasks, however long they took.
- Remove tasks that are not rule-based. Some of these need a better process, not a tool.
- Score the rest. Use the four-question grid in the business process automation guide: how often it runs, how clear the rules are, whether the data is in one place and how cheap a mistake is.
- Do the payback arithmetic for the top two or three: monthly time saved, minus upkeep, against build time.
- Pick one and build it. Then run the same log again after four to eight weeks to see whether the hours really moved.
The same method works in any industry. In recruitment, the recurring tasks that usually surface are record creation, scheduling and status updates, which are covered in twelve processes you can automate in a recruitment agency.
Common mistakes
- Logging from memory. Reconstructed logs underestimate small interruptions and over-round everything.
- Logging only the tasks you dislike. Log everything for the week, or the totals mean nothing.
- Letting categories drift. If "status update" is sometimes "reporting" and sometimes "admin", the totals split and the biggest task hides.
- Counting one-offs. A single long task is not a recurring cost.
- Using the log to monitor people. It will stop being honest. Use it to find processes to fix.
- Auditing once and stopping. Repeat after a few months, because work patterns change.
Next step
Set up the log today and run it for the next five working days. If you would like a second pair of eyes on the results and what is worth automating, you can book a free audit. For a fuller treatment of one industry, see what to automate first and what to keep human in a recruitment agency, or browse the rest of the blog.
Sources
- McKinsey Global Institute, The social economy: Unlocking value and productivity through social technologies (July 2012)
- World Economic Forum, These 3 charts from Microsoft show why technology is making your job harder (reporting on the Microsoft Work Trend Index 2023)
- Gallup, Too Many Interruptions at Work? (reporting on Gloria Mark's research)