Machine Hour Rate in India: Count the Hours You Never Sold

Your machine hour rate (MHR) is the cost of running one machine for one hour — total machine-related cost divided by productive hours. The number almost everyone gets wrong is the divisor. Set it on the hours you assume the machine runs and you under-recover cost on every job; set it on the hours the spindle actually cuts and the rate can double. The machine didn't get more expensive — you were quietly giving away the hours you never sold.

By the Mama Editorial Team. Most Indian job shops and auto-component units carry a machine hour rate in their costing sheet, quote against it, and never revisit it. This piece is about the one input that decides whether that rate protects your margin or leaks it: loading. We work it in rupees for one CNC.

Key takeaways

• MHR = total annual machine cost ÷ productive machine hours. The numerator is easy; the denominator is where the money hides.
• Normative (budgeted) hours flatter the rate. If you divide by optimistic "normal capacity" but the spindle only cuts 40% of planned time, you under-absorb overhead on every job.
• Real loading is usually far below the sheet. Industry estimates put typical small-to-mid CNC shops at ~25–50% spindle utilisation (MachineMetrics, JITbase — see benchmarks below).
• The gap is cash. In the worked example, the same machine costs ₹742/hour on paper and ₹1,576/hour in reality — about ₹834/hour quietly unrecovered.
• You can't fix the divisor you don't measure. The fastest win is not a cheaper machine; it's an honest, timestamped record of cutting hours.

What machine hour rate actually is

MHR converts the lumpy annual cost of owning and running a machine into a per-hour figure you can load onto a job. Add up everything the machine costs in a year — then divide by the hours it will be productive:

MHR = (depreciation + interest + power + operator + maintenance + tooling/consumables + allocated overhead) ÷ productive machine hours.

The numerator is standard costing and most owners get it roughly right. The denominator is a choice, and it is where the error lives. Pick a big divisor and the rate looks competitive; pick an honest one and it looks expensive. The market does not care which you picked — it only pays for the hours the spindle was actually earning.

The denominator decides everything

There are three different "hours" in play, and conflating them is the whole problem:

Standard costing deliberately uses normal capacity — a budgeted, achievable level — so that the rate is stable and under-recovery shows up as a visible variance rather than being hidden. That is sound. The failure in practice is that the "normal capacity" people plug in is really nameplate dressed up, and the under-recovery variance is never looked at. The machine is assumed to cut 1,900 hours; it cuts 900; nobody reconciles the 1,000-hour hole.

A worked ₹ example (one mid-size CNC)

A single CNC machining centre, single shift. Figures are indicative October 2026 estimates — plug in your own.

Now the only question that matters: divide by what?

Planned time, single shift = 8 h × 26 days × 12 months ≈ 2,496 h; net of breaks and planned stops, call it 2,300 planned hours.

• On normative loading (≈85% of planned, the convention a cost accountant would use): ~1,955 productive hours. ₹14,50,000 ÷ 1,955 = ₹742/hour.
• On real loading from the floor (spindle actually cutting 40% of planned time — mid-range for this kind of shop): ~920 cutting hours. ₹14,50,000 ÷ 920 = ₹1,576/hour.

If you quote jobs at ₹742 while the machine's true absorbed cost is ₹1,576, you under-recover roughly ₹834 for every hour the spindle turns. Across 920 cutting hours that is about ₹7.7 lakh a year on one machine that never shows up as a loss — it just quietly erodes the margin on every job you win. Multiply across a ten-machine shop and it is a seven-figure line nobody can point to.

Two honest caveats. First, not all of the gap is "lost" — setup, load/unload and inspection are legitimate non-cutting time, and a shop that bills by cutting minute should load those into the rate deliberately, not pretend they don't exist. Second, the fix is rarely to raise the quoted rate (the market may not bear it); it is to move the divisor — turn more planned hours into cutting hours. That is a utilisation problem, and it is measurable. We go deeper on the bands in Spindle Utilisation: What 25%, 45% and 70% Mean on an Indian Floor.

Why real loading is almost always lower than the sheet

Owners estimate their own utilisation at 70–80%; measured studies repeatedly land far below. MachineMetrics, which instruments large numbers of machines, has long argued that average discrete-manufacturing utilisation sits near the low-to-mid 20s to 30s percent, against an owner's gut of 70%+ (MachineMetrics, machinemetrics.com). JITbase and Modern Machine Shop benchmarking similarly place typical job-shop spindle-cutting time well under half of attended time. Treat the exact figure as an industry estimate, not a law — your number depends on part mix, batch size and changeover frequency — but treat the direction as reliable: the sheet assumption is optimistic, usually by a wide margin.

The hours vanish into things no costing sheet tracks: setups, waiting for material or the crane, tool changes, chasing a sign-off, micro-stops, and the gap between "machine switched on" and "spindle cutting." The same leakage that wrecks OEE wrecks your MHR divisor — the same hours counted two ways. And when the machine sits idle but energized, you still pay the sanctioned-kVA demand charge and standing costs: the rupees in the numerator keep accruing while the hours in the denominator don't.

How to get the divisor right

1. Separate the three "hours." Write down nameplate, your budgeted normal capacity, and — the one you probably don't have — measured cutting hours.
2. Measure cutting hours for two weeks on your busiest machine. Not "was it switched on" — was the spindle cutting.
3. Recompute MHR on the real number and compare to your quoting rate. The gap is your annual under-recovery.
4. Decide what the rate must cover — if you bill by cutting minute, deliberately load setup and idle into it; don't let them leak.
5. Then attack the divisor, not the rate. Cutting changeover and wait time adds cutting hours, which lowers a true MHR without touching your price — the only free margin in the building. See auto-component shops passing IATF while they do it.

Measuring cutting hours without re-wiring every machine

The hard part is step 2 on an old mixed fleet — Fanuc, Mitsubishi, manual lathes — where there is no MES and no clean signal to tap. You can clamp a current sensor on each spindle drive, you can have operators log it, or you can let a camera already pointed at the floor watch spindle motion and chip flow and timestamp every start and stop. Each has a different cost and a different level of trust; we compare all three, in rupees, in Tracking Old CNCs Without a PLC.

This is the gap Mama is built to close. Record a short phone walk of your shop and send it over; Mama reads the feeds — your existing cameras over RTSP/ONVIF with no replacement, or cameras we add — and every morning it sends the owner a plain-language WhatsApp note: where the hours leaked yesterday and what they were worth in rupees. It watches machines, not people — the point is the idle spindle, never a ranking of operators. Send a floor video and you get back where the hours go, a camera plan for the blind spots, and a proposal. The first thing that usually changes is the divisor in this very calculation.

FAQ

What is the machine hour rate? It is the cost of operating one machine for one hour: total annual machine-related cost (depreciation, interest, power, operator, maintenance, tooling and allocated overhead) divided by the machine's productive hours. It lets you load machine cost onto a job or a quote.

How do you calculate machine hour rate in India? Add the machine's annual costs, then divide by its productive hours. The arithmetic is simple; the judgment is the divisor — use realistic, measured productive hours rather than nameplate or optimistic "normal capacity," or you will under-recover cost on every job.

Why is my machine hour rate too low? Almost always because the hours in the denominator are too high. If the sheet assumes the spindle cuts ~1,900 hours a year but it really cuts ~900, the true cost per cutting hour is roughly double the rate you are quoting.

Is machine hour rate the same as OEE? No, but they share a denominator. OEE measures the lost hours as a percentage; MHR converts the remaining productive hours into a cost. Low utilisation shows up as a poor OEE and as an under-stated MHR — the same hole, counted two ways.