- Why headcount is the wrong test
- Seven checkpoints for Excel attendance
- Three rules to agree before moving
- Starting fresh without migrating old data
The short answer
Deciding whether to drop the spreadsheet is not about headcount. It is about how many times a person has to touch the data during each month-end close. If nothing gets re-typed or corrected and the close finishes in one pass, a spreadsheet is fine. If a ten-person company repeats the same corrections every month, moving off it will pay back quickly.
When a spreadsheet is enough
Everyone starts and finishes at the same time, there is one location, staff are on monthly salary rather than hourly pay, and paid leave is taken in predictable blocks. In a workplace like that, the close means counting working days and handing the total to payroll. You do not even need formulas, and the effort is close to a paper sign-in sheet. Forcing that situation into a system only adds the cost of switching.
Spreadsheets get painful when exceptions pile up. Someone steps out mid-shift, someone goes straight from home to a customer site, someone's hourly rate changes mid-month, someone splits time between two sites. Each exception adds a column, and every new column adds another place where a formula has to be fixed.
Seven checkpoints
Look at these one at a time instead of at headcount. If three or more of them involve manual work every month, it is time to consider a system.
Clock-in
Ask who enters the start and end times, and when. If people enter them as they happen, the risk is small. If they reconstruct the whole month from memory on the last day, that is a declaration, not a record. Japan's Ministry of Health, Labour and Welfare guideline on properly grasping working hours treats two methods as the principle: the employer confirms and records the times directly, or the record is based on objective data such as time cards, IC cards, or PC usage logs. Where self-reporting is unavoidable, the employer is expected to investigate and correct the figures when reported hours differ noticeably from reality. End-of-month batch entry is the shape that produces that corrective work every single month.
Edit history
Clock-in mistakes are unavoidable. The problem is not the correction itself but the fact that nobody can tell who changed what, and when. Overwriting a spreadsheet cell erases the previous value, so you cannot distinguish a supervisor adjusting a report from an employee rewriting their own hours later.
Month-end close
If the file is handed over twice or more before payroll sees it, versions will diverge somewhere along the way. When "which file is the latest one?" is a monthly question, the close is not yet a process.
Breaks
If you deduct a flat 45 minutes or one hour, check that the deduction matches what actually happens. In workplaces where people step out or travel between sites, a flat deduction drifts away from reality.
Paid leave
Tracking remaining days in a spreadsheet means each row carries its own formula, because grant dates differ per person. Companies that grant leave based on each employee's start date have the most fragile column here. Rebuilding the sheet for a new year also turns the carry-over of remaining days into manual work.
Multiple sites
When each site keeps its own file, someone consolidates them every month. If the layouts differ even slightly, that person fixes the layouts before totalling anything. This is where moving to a system tends to show the clearest gain.
Payroll handoff
If the totals are re-typed into payroll software, you are copying every employee's figures once a month. Whether the data can move as a file such as CSV makes a considerable difference to the workload on the close date.
What this looks like in practice
Picture a company with an office and a warehouse, eighteen people including part-time staff. The warehouse runs shifts, and people join or leave mid-month. The administrator distributes sheets at the start of the month, collects them at the end, reformats the warehouse ones, counts late-night hours by hand, and keys eighteen sets of figures into payroll. That is five manual steps: distribute, collect, reformat, count, re-type.
If that company shrank to eight people, all five steps would remain, because shift work and two locations have not changed. That is what "headcount alone is not the answer" means. The reverse also holds: a thirty-person company with one location and one work pattern may run comfortably on a spreadsheet.
Decide these before you move
Write down three things before migrating.
- Who records the time — the employee or a manager
- Who is allowed to correct a record
- What happens when something has to be fixed after the close
Skip the third and post-close corrections will be handled outside the system, in email and chat, and the spreadsheet will quietly come back.
Trying to migrate every past spreadsheet is what stalls most migrations. Keep the history as files and start the new process from the switchover month.
Limits
A system does not make working-time management correct by itself. If the clock-in data does not match reality, the system will total inaccurate data accurately. Retention periods and the range of documents covered can also change over time, so check the current requirements in the Ministry of Health, Labour and Welfare's own guidance.
Kintai, APOUL's attendance product, is planned for release and not yet available. A tool helps with recording and totalling; the rules about what counts as a record still have to be set by the company.
Summary
- Decide by how often a person touches the data at close, not by headcount.
- Check clock-in, edit history, close, breaks, paid leave, sites, and payroll handoff.
- Agree on who records, who corrects, and how post-close fixes work.
- Leave historical data as files and start fresh from the switchover month.
- Confirm detailed legal requirements with the ministry's own guidance.
APOUL KINTAI