Why Are My Machines Idle When the Job Sheet Is Full? Four Causes, Priced in ₹
Why are my machines idle when the job sheet is full? Because a job on the sheet is not a job at the machine. As of 17 Oct 2026, we sort such stops into four waits: no material, no tool or fixture, no decision, or no operator. In our assumed 20-machine shop they cost about ₹8.1 lakh a month at ₹800 per machine-hour.
By the Mama Editorial Team. Written for owners of CNC and fabrication job shops who see a full planning board and still count machines standing. Last checked: 17 Oct 2026. Every ₹ figure in the worked example is a stated assumption, not a benchmark.
Key takeaways
- A full job sheet proves demand, not readiness. A machine runs only when material, tooling, a released decision and an operator reach it together.
- Four waits, four owners. Material belongs to stores and purchase, tooling to the tool room, decisions to QC, engineering or you, and operators to the supervisor.
- Each wait has a visible floor sign. Empty bin beside a set machine, a setter walking with a holder, a part lying on a table, a ready machine with nobody at it.
- Price idle hours at your machine-hour rate. In our worked example, 234 waiting hours a week across 20 machines come to about ₹97 lakh a year, which equals roughly 2.4 machines bought and left switched off.
- Measure for two weeks with a four-letter tag (M, T, D, O) before you spend on anything. The biggest letter tells you whose desk to visit first.
What this page covers and what it leaves out
This page covers one symptom: orders are waiting, yet machines are not cutting. Breakdowns, planned setups and slow cycles need different fixes and are left out. If every machine is running but output is still short, read Output 20% Below Plan While Every Machine Is Running. For how much of the shift spindles typically cut, see Spindle Utilisation: What 25%, 45% and 70% Mean on an Indian Floor.
Why "the order book is full" does not keep machines busy
ACMA's half-year review, released on 14 January 2026, put auto component industry turnover at ₹3.56 lakh crore for April to September 2025, up 6.8% on a year earlier (ACMA H1 FY26 press release). That is industry context, not a figure about your shop.
The same ACMA release named "the limited availability of critical materials such as rare-earth magnets" and "raw-material price volatility" among the industry's headwinds (ACMA, 14 Jan 2026). Demand and inputs do not move together; that is how a full job sheet and an idle machine sit side by side.
ACMA's H1 FY26 review gives no figure for machine idle time or capacity utilisation at component makers (checked 17 Oct 2026).
Cause 1: No material
The machine is set and the operator is there, but the bar stock, forging, casting or the part from the previous operation has not arrived.
What you see on the floor: an empty bin or pallet beside a machine that was running an hour ago. The operator is cleaning, or standing near the stores counter. Often an outsourced step such as heat treatment or plating came back late.
Where it really starts: the planning board shows the job, but nobody checked that material was physically in the shop before the machine was booked. A forging lot is short and nobody tells the planner until the bin is empty.
Who owns it: stores and purchase, with the planner.
First fix: make "material in the shop, counted" a condition for booking a job on a machine. Stage a kitted trolley for the next job before the current one ends.
Cause 2: No tool, fixture or gauge
Material is there, but the right insert grade, boring bar, soft jaws, fixture or thread gauge is not.
What you see on the floor: a setter walking between two machines carrying the same holder. A fixture sitting at the tool room "for a small correction". Two machines that run the same family of parts taking turns, because there is only one set of jaws or one go/no-go gauge for that thread.
Where it really starts: tooling is bought against breakdowns, not against the job sheet. Inserts are reordered when the box is empty. Fixtures are shared because a second one looked expensive; the idle hours never show up on a bill.
Who owns it: the tool room and whoever approves tooling purchases.
First fix: keep a tooling list per part number and check it against next week's jobs every Friday. A missing holder (assume it costs ₹3,000) that idles a machine for one 8-hour shift costs ₹6,400 at the ₹800 machine-hour rate used below, more than twice its price.
Cause 3: No decision
Material, tool and operator are all present, but somebody has to say "go" and has not yet said it.
What you see on the floor: a set machine with a part lying on the table next to it. People standing around a drawing. A supervisor on the phone. The setter waiting for QC, for engineering, or for you.
Where it really starts: a frequent decision is first-piece approval after a setup, and it behaves like a queue at the inspector. We cover that queue and its fixes in Waiting for First-Piece Approval, so we do not repeat it here. Other decisions: a drawing query to the customer, a material deviation, rework-or-scrap on a rejected lot, and "which job next" when the planner is absent.
Who owns it: QC, engineering and, more often than owners like to admit, the owner. If the floor waits for you to answer a priority question, the machine idles at your machine-hour rate while your phone is busy.
First fix: write down who may decide what without escalation, for example "the supervisor may resequence jobs within one customer".
Cause 4: No operator
The machine is ready, the job is released, and nobody is at it.
What you see on the floor: a machine with its door open, lights on, nothing loaded, and no person within a few metres. The operator was moved to another machine, is still at the canteen, or did not come in.
Where it really starts: one operator is assigned to two machines on paper and to one in practice. Absence is planned for on a holiday, not on an ordinary Monday. In a study of four garment factories in Karnataka, 10 to 11% of workers were absent on a typical day, and nearly all of that absence was unauthorised (Adhvaryu et al., Ideas for India, 8 Sep 2021). That is a different sector from yours, so treat it as a reminder that daily absence is normal, not as your figure.
Who owns it: the shift supervisor and whoever approves manning.
First fix: decide in advance which machines stop first when people are short, so the idle hours fall on low-priority jobs. Treat it as a manning question, never a judgement on one person.
Worked ₹ example: one week in a 20-machine job shop
This example is entirely assumed; it shows the method, not your loss.
Assume a 20-machine CNC job shop (14 turning centres, 6 VMCs) in Rajkot, Pune or Coimbatore, running two 8-hour shifts, six days a week. That gives 96 hours per machine and 1,920 available machine-hours a week. Assume a blended machine-hour rate of ₹800, meaning the cost you planned to recover for every hour each machine cuts. If you do not have a rate you trust, work one out with the machine hour rate calculator. Why the divisor in that rate matters is explained in Machine Hour Rate in India: Count the Hours You Never Sold.
Now assume two weeks of idle tagging produced these average waiting hours per week, all on days when the job sheet was full:
| Cause (tag) | Idle hours per week (assumed) | Share of available hours (%) | ₹ per week at ₹800/hour | ₹ per month (× 52 ÷ 12) | ₹ per year (× 52) |
|---|---|---|---|---|---|
| No material (M) | 72 | 3.8 | 57,600 | 2,49,600 | 29,95,200 |
| No tool, fixture or gauge (T) | 48 | 2.5 | 38,400 | 1,66,400 | 19,96,800 |
| No decision (D) | 60 | 3.1 | 48,000 | 2,08,000 | 24,96,000 |
| No operator (O) | 54 | 2.8 | 43,200 | 1,87,200 | 22,46,400 |
| Total | 234 | 12.2 | 1,87,200 | 8,11,200 | 97,34,400 |
How to read it. 234 waiting hours a week is 12,168 machine-hours a year. One machine in this shop is available for 4,992 hours a year (96 × 52). So the waiting equals about 2.4 machines that you bought, powered and staffed, and then left off for the whole year.
Why this is a floor. The machine-hour rate counts only the cost you failed to recover. With a full job sheet, the job that did not run also delays a delivery or forces overtime later; neither is in this table.
What the table does not say. Material is not always the biggest cause. The point is that each cause gets its own number, so each fix gets its own owner.
How to measure the four causes in two weeks
The rule: whenever a machine stands for more than 10 minutes while it has a job assigned, the operator or supervisor writes one letter on a sheet hung on the machine: M for material, T for tool, fixture or gauge, D for decision, O for operator. Anything else (breakdown, planned setup, power cut) gets X and is left out of this exercise.
The sheet: machine, date, start, end, letter, and one word of detail ("forging", "insert", "FPA").
The review: every Saturday, add up hours by letter. After two weeks, multiply by your machine-hour rate. Take the largest letter to its owner, with the number, and agree one change.
Expect two traps: tags drift toward the least embarrassing letter, such as M, which blames a supplier, and the sheet stops being filled in week two unless someone collects it daily.
What cameras can and cannot tell you here
A camera covering the machine rows can show, without anyone filling a sheet:
- when each machine stopped and restarted;
- whether anyone was at a standing machine, counted in aggregate, not by name;
- an empty bin beside a set machine, or a part waiting on a table;
- how many machines stood at the same hour.
It cannot tell you:
- What is on the job sheet. Whether a machine had a job assigned lives in your planning board or ERP, not in the frame.
- Whether the insert or fixture exists. Video sees a setter walking, not the tool-room stock register.
- What decision is pending, or who must make it. A part on a table looks the same whether it waits for QC, a drawing query or your call.
- Why someone is absent. That is attendance and HR, and it should stay there.
Video and the four-letter tag work best together: the camera gives honest start and end times, and the tag gives the reason. Keep the review about machines and causes, never about ranking individuals. Footage of recognisable workers is personal data under s.2(t) of the DPDP Act, 2023; tell workers in writing what is recorded and why (see the worker CCTV notice under DPDP).
Where Mama fits
Mama reads the cameras a factory already has over RTSP/ONVIF, or a few added ones, and every morning sends the owner a short plain-language note on which machines stood the day before, for how long, and whether anyone was at them. It watches machines and flows and counts people only in aggregate; it does not read your job sheet or name anyone, so your own tags and records still supply the reason. To see what that would look like on your floor, Leave an email or phone in the form below, and Mama will get back to you with a camera plan for any blind spots.
FAQ
Why are machines idle when there are plenty of orders? Because an order still needs material, tooling, a released decision and an operator at the machine at the same moment. Miss one and the machine waits.
Which of the four causes is usually the biggest? There is no published figure for Indian job shops; ACMA's H1 FY26 review, for example, gives no idle-time split (checked 17 Oct 2026). Tag your own stops for two weeks: in our assumed example material came first, but your own split can look quite different.
How do I put a rupee value on idle machine hours? Multiply idle hours by your machine-hour rate, which is total annual machine cost divided by productive hours. With a full job sheet, treat the result as the minimum loss, because late deliveries and later overtime come on top.
Is waiting for first-piece approval counted as idle time? Yes, in this scheme it is a "decision" wait (D), because the machine is ready and only a sign-off is missing. The queue at the inspector and how to shorten it are covered in our first-piece approval article.
Should I hire more operators if many machines stand with nobody at them? Not before you check the other three letters. When machines wait for material, tooling or decisions, operators drift away and the stop looks like "no operator" when it is not.