Man-Month Effort Calculator

There is no universal answer, which is exactly why estimates disagree. In Korea, KOSA derives its published daily wage by dividing the monthly wage by 20.5 days for the 2026 figures, so 20.5 is the number consistent with the official rate table. The widely repeated “1 M/M = 21 days” is not a KOSA rule — it is a rounded rule of thumb, and 22 days is another common convention that simply counts weekdays. Japan instead defines the month by hours: 160 h (8 h × 20 days). Note that KOSA's divisor moves a little every year (20.8 in 2022, 20.6 in 2024, 20.5 in 2026), so a man month calculator that hard-codes one constant silently goes stale. This tool keeps all of them as presets with the source year on the label, and you can set a custom hours/day and days/month pair.

Everything normalises through hours: man-days = man-hours ÷ hours-per-day and man-months = man-hours ÷ (hours-per-day × days-per-month). With the KOSA 2026 preset (8 h/day, 20.5 days/month) one man-month is 164 h, so a 1,312-hour project is 1,312 ÷ 8 = 164 man-days = 1,312 ÷ 164 = 8 M/M. Switch to Japan's 160 h and the same 1,312 hours becomes 8.2 M/M — the work did not change, only the divisor did. Paste a whole WBS and this man day effort estimate runs that arithmetic per row, then rolls the totals up so you never rebuild the sheet by hand.

KOSA (Korea Software Industry Association) surveys software engineer wages and publishes the table each December for the following year — the 2026 figures were published in December 2025. Crucially it is a loaded cost, not take-home pay: it includes base pay, allowances, bonuses, the severance provision, and the employer's share of the four major insurances. That is why it looks high next to a salary, and why it is the right basis for a quote. Because it is published once a year, this tool keeps the constant in one table and only that number changes annually. The built-in figure is KOSA's single all-roles average; per-role numbers are yours to enter from the survey, since publishing role-level figures we cannot verify would be worse than asking.

A lot — and it is the most common source of argument in Japanese contracts. The same 1,600 man-hours is 11.43 M/M at a 140 h divisor, 10.0 M/M at 160 h, and 8.89 M/M at 180 h: a 29% swing in the headline number for identical work. When the Japan preset is selected, this calculator shows all three side by side so you can see how exposed the estimate is before you sign the settlement range. Calculating the actual over/under settlement payment (上下割 / 中間割) is deliberately out of scope: it needs recorded actual hours and per-person monthly contracts, which this planning-effort tool does not ask for and should not pretend to know.

No — and you do not have to take our word for it. There is no server and no account: the parsing, the man-month rollup by role, and the cost maths all run in JavaScript in your tab. Verify it yourself: open DevTools (F12) → Network tab, paste your WBS and press Calculate — you will see zero requests, and you can disconnect from the internet entirely and the tool still works. Only your reusable settings (rate table, chosen constant, column mapping) are kept in this browser's localStorage; task names, effort and costs are never stored or transmitted. That is the point: quote cost data is usually the most sensitive thing a PM handles.