# The five-day method

The Hour Ledger method for measuring one back-office task. Published free.
Copy it, print it, run it with a spreadsheet, a free tracker or a paper sheet
on the desk. Nothing on this page is charged for.

Version 1, 7 September 2026. Praxis by Superalignment.

## 1. Write the task in one sentence

The sentence needs a start event and an end event. Without both, you will
measure a category rather than a task, and the week will produce a number that
no one can act on.

Passes:

- "Enter supplier invoices into the accounting system and match them to a
  purchase order. Starts when the invoice arrives in the shared accounts inbox.
  Ends when it is posted and matched."
- "Build the weekly service schedule. Starts when Friday's job list is final.
  Ends when every technician has Monday's route."

Fails:

- "Accounts payable." That is a department, not a task.
- "Admin." That is a mood.
- "Fixing supplier problems." No end event, so no two people will agree when a
  block of work stopped.

## 2. Name the people and the rate

List the two or three people who actually touch the task, plus anyone who gets
pulled in on exceptions. Write one loaded hourly rate per person: wage plus
payroll taxes plus benefits, divided by paid hours. Keep the rates to yourself.
The people marking the sheet do not need to see them, and the study is easier to
run when they do not.

## 3. Tell the people what is being measured

Say it out loud before the week starts: the study is about the task, not about
them. A tally that people fill in to look busy, or to look efficient, measures
neither. If someone believes the number will be used against them, stop and fix
that first, because you will not get an honest week out of them.

## 4. Mark the sheet as the work happens

Five consecutive working days. Fifteen-minute blocks. Each block gets a cause
code. Marks go on the sheet at the moment the work happens, not from memory at
the end of the day, and not from memory on Friday. Reconstructed hours are the
weakest form of this evidence and they are the reason most time audits end in an
argument.

Six cause codes, adapted to your task. The set below is for supplier invoice
entry:

- A. Chasing a missing or wrong purchase order number
- B. Re-keying from a PDF or paper invoice
- C. Price or quantity mismatch against the purchase order
- D. Finding the person who can approve it
- E. Duplicate, resent or misdirected invoice
- F. Other, named on the line

Rules for the codes: six at most, one code per block, and F never exceeds a
quarter of the week without you writing down what is in it. An "other" bucket
that swallows the week is how a study turns into a shrug.

## 5. Check the day, every evening

One person owns the sheet. Each evening they confirm that every named person
either logged something or marked the day as none. A day marked "none" is data.
A day left blank is a hole.

## 6. Score the week

Add the daily rows. You want four figures:

- Hours on the task per person for the logged days
- The split by cause code
- The cost line: hours multiplied by each person's loaded rate
- The largest single cause, as a share of the logged hours

Do not annualize inside the study. If you want a year, multiply afterward by
the number of weeks you actually work, and say out loud that you did it.

## 7. The refusal rule

- Fewer than three of the five days logged: the week is unusable. Do not
  produce a number from it. Rerun the week, and pick a week without a holiday,
  a shutdown or an install push in it.
- A person logged nothing all week: report their row as unmeasured, exclude it
  from the total, and label the total a floor rather than an estimate.
- A day nobody logged: exclude the day. Do not estimate it from the other days.

A floor you can defend beats an estimate you cannot.

## 8. Write the decision down

The week ends with one of three sentences, and the reason:

- Fix it internally.
- Buy something.
- Leave it alone.

"Leave it alone" is a real result. Plenty of tasks that feel enormous turn out
to cost four hours a week, and knowing that is worth the week it took to find
out.

## What this method is not

It is not a payroll, wage-hour or compliance instrument, and it is not evidence
for any employment matter. It produces no benchmark against other companies,
because we have none to offer. It makes no claim about what your number will be.