Manufacturing work-order labor variance software
Your ERP tells you a work order ran over after it closes. Takum reads the daily hours and flags the drift while the order is still open.
30-50%
McKinsey measured what happens when high-mix, low-volume manufacturers build real-time visibility and daily performance planning into standard work: productivity typically improves 30 to 50 percent within eight to 12 weeks. This page walks through the problem those plants fixed, and how.
/ 01 · The problem
Variance found at close is history, not control.
Work-order labor variance is the gap between the hours a routing says a job should take and the hours your people actually charge to it. Every manufacturer that runs work orders has it. The question is when you find out.
Most plants find out at month-end close. By then the order is finished, the crew has moved on, and the variance report is a history lesson. You can argue about the standard. You can code the loss to a bucket. What you cannot do is change how the job was run, because the job is over.
The useful moment is earlier: day 2 or day 3 of an open order, when the hours are drifting past pace and a supervisor can still re-sequence the work, move people, or find out the standard was wrong before it burns through the rest of the schedule.
/ 02 · The root cause
The overrun is not sudden. It is unmeasured.
A job rarely blows through its standard in one bad shift. It drifts. A setup runs long, a part fights back, a crew works short one person. That drift shows up in the daily hours from the first day it starts.
The root cause is measurement timing. In shops without shop-floor data capture, actual hours ride payroll's calendar and land weekly, sometimes biweekly. Variance gets computed at close, because that is when accounting needs it. And pace, the percent of hours consumed against the percent of work complete, is rarely computed at all. The one number that would send a supervisor to walk the job today does not exist in most plants.
This is not a niche failure. In Accenture's 2023 survey of 1,230 senior operations executives, only 11 percent of companies had near real-time alerting, and 78 percent needed at least a week to fully understand the impact when something goes wrong.
None of this is new. Harvard Business School's Robert Kaplan, with H. Thomas Johnson, put it on the first page of their 1987 book on management accounting:
"Today's management accounting information... is too late, too aggregated, and too distorted to be relevant for managers' planning and control decisions."
Johnson & Kaplan, Relevance Lost, Harvard Business School Press, 1987
That was written almost four decades ago. In most plants, the reporting cycle it describes has not changed. Stale standards make it worse: when a routing has not been updated since the process changed, the variance report reads as noise, and people stop trusting it. A number nobody trusts is a number nobody acts on.
/ 03 · The solution
How labor variance software works in practice.
Feed
A daily file of hours
Each day, hours by employee and work order come over from your time charging. A flat file or an export from the ERP is enough. No new time clocks, no shop-floor rollout.
Compare
Hours against the standard
Hours to date get compared against the standard for that routing, at the point the order is in its life. An order at 40 percent of its hours with 20 percent of the work done is a different fact than the same total at 90 percent done.
Flag
While the order is open
Orders trending past standard get flagged. The supervisor sees which order, which day the drift started, and how far past pace it is running.
One honest caveat: a flag is a question, not a verdict. Sometimes the job is genuinely in trouble. Sometimes the standard is wrong. Both answers are worth having while there is time to use them.
Built to catch it early, not explain it late.
Takum is cost-control software that identifies work-order labor overruns and unsupported vendor charges while there is still time to act.
For a manufacturer, that means Takum takes the daily hours feed described above, learns what normal pace looks like for each routing, and surfaces the orders that are drifting, in time for your team to step in. It sits beside your ERP rather than replacing it. Your standards, your work orders, and your time charging stay exactly where they are.
Takum does not promise that every flag saves money. People still make the call on the floor. What it removes is the excuse that nobody could have seen the overrun coming, because with a daily feed, someone can.
/ 04 · Illustrative example
What a catch on day 4 is worth.
The numbers below are invented round numbers chosen to show the math.
Caught on day 4, that order is a conversation: re-sequence the remaining operations, move people, or check whether the standard itself is wrong. Caught at close, it is a variance line and an awkward meeting.
An operator taught us the principle behind this page: what gets measured gets managed. Kaplan and Norton opened their 1992 Harvard Business Review article on measurement with the same idea: "What you measure is what you get." The day-4 flag is that principle applied to labor hours. Nothing about it is magic. It is a number someone sees in time to act on it.
The $65 loaded rate is an assumption, not a measurement. BLS puts total employer compensation for manufacturing workers at $48.27 per hour worked as of March 2026, wages plus benefits; loaded rates that add overhead and indirect costs run higher, and $65 is our round working number. Swap in your own; the math is the point, not the rate.
/ 05 · Fit
Who this is for, and who it is not.
A good fit if
- You run work orders with routings and labor standards, and hours get charged to them by employee.
- You can produce a daily export of those hours, or your ERP already can.
- Labor variance shows up at month-end, and by then nobody can explain it or do anything about it.
Not a fit if
- Labor is not charged to work orders. With nothing to compare against a standard, there is nothing to flag.
- You only want job costing after close. Your ERP already does that, and Takum would not add much.
- You cannot produce a regular hours feed. Takum cannot flag what it cannot see.
See your variance while you can still act on it.
Request a demo and see how a daily hours feed turns into flags your supervisors can act on the same day.