UPH, UPPH and Line Balancing with Indian Wage Maths

By the Mama Editorial Team — we scope camera projects for Indian factories. Last verified: 2 October 2026, against standard industrial-engineering definitions and the wage/productivity sources cited inline.

UPH (Units Per Hour) is how fast a line produces; UPPH (Units Per Person Per Hour) is how fast it produces per operator on it. UPH tells you throughput, UPPH tells you labour productivity — and on a manual line, UPPH is set almost entirely by the slowest bench. Fix the bottleneck and rebalance, and you raise both numbers without hiring anyone. This page works it in rupees.

For a mid-size Indian contract manufacturer — electronics assembly, box build, wire harness, sub-assembly — labour is a cost you control directly, unlike copper or components. The two numbers that govern it are UPH and UPPH, and most plants track UPH but never calculate UPPH, which is the one that actually appears in the P&L. Here is both, with a worked example and the wage maths that makes it matter in India.

Key points

• UPH = line output per hour. UPPH = UPH ÷ number of direct operators on the line. UPPH is the labour-productivity number.
• The bottleneck sets the line. A line runs at the speed of its slowest station, not its average — so one slow bench drags every operator's UPPH down.
• Line-balancing efficiency = (sum of all station work times) ÷ (number of stations × the bottleneck station's time). The gap from 100% is paid idle time.
• Rebalancing is usually free. Moving work from the bottleneck to a starved neighbour raises UPH with the same headcount.
• India's wage advantage is real but smaller than it looks once you adjust for productivity (unit labour cost, industry estimate) — which is exactly why UPPH matters more here, not less.
• Video measures cycle and idle, not quality. A camera can time each station and find the bottleneck; it cannot judge a solder joint or a test pass.

UPH vs UPPH: the definitions, and why the second one matters

UPH (Units Per Hour) is throughput: good units leaving the line in an hour. It answers "how fast is the line?"

UPPH (Units Per Person Per Hour) divides that by the number of direct operators:

UPPH = UPH ÷ number of direct operators

It answers the question an owner actually pays for: "how much output am I getting per pair of hands?" A line can raise UPH by throwing more people at it and still see UPPH fall — more output, but less per worker, and a higher labour cost per unit. That is why UPPH, not UPH, is the number to defend.

The bottleneck decides everything

On any line of sequential stations, the output rate is capped by the slowest station — the bottleneck. If station 4 takes 60 seconds while every other station takes 45, the whole line produces one unit every 60 seconds no matter how fast the others work. The faster stations simply wait. That paid waiting is the cost of imbalance.

Line-balancing efficiency quantifies it:

Balance efficiency = (Σ station work times) ÷ (number of stations × bottleneck time)

The shortfall from 100% is balance delay — idle operator time you are paying for. This is textbook industrial engineering, and it is the lever most manual Indian lines never pull, because nobody is timing each station. For the broader "see the bottleneck in time" problem across a plant, see why your shipments slip.

Worked example, in rupees

A manual assembly line, 8 stations, 8 direct operators, single 8-hour shift. Station times, in seconds:

Station 4 at 60 s is the bottleneck, so the line produces one unit every 60 seconds.

UPH = 3,600 ÷ 60 = 60 units/hour (at ideal; real output is lower after breaks, micro-stops and waiting).

UPPH = 60 ÷ 8 = 7.5 units per person per hour.

Balance efficiency = (40+45+42+60+44+41+43+45) ÷ (8 × 60) = 360 ÷ 480 = 75%. A quarter of your paid operator-time is balance delay.

Now the money. Take an illustrative fully-loaded operator cost of ₹18,000/month, ≈ ₹86.5 per operator-hour (26 days × 8 hours). Eight operators cost ₹692/hour.

Rebalance. Move ~15 seconds of work off station 4 onto the lightly loaded stations 1 and 6. New bottleneck ≈ 48 s:

Same eight people, same wages, and labour cost per unit falls from ₹11.5 to ₹9.2 — about 20% — purely from moving work between benches. Across a shift that is 120 extra units at no extra labour cost. Nobody was hired; a stopwatch and a reshuffle did it. The numbers are illustrative — plug in your own station times and wage — but the mechanism is exact.

The India wage angle — and the caveat that matters

The usual story is that Indian labour is cheap, so productivity per worker matters less. That is half-right, and the missing half is what makes UPPH more important here, not less.

India's manufacturing wages are a fraction of China's — a real, large gap (industry estimate; exact multiples vary by source and sector). But China's manufacturing labour productivity — value added per worker — is also substantially higher, so the gap in unit labour cost (wage per unit of output) is much smaller than the raw wage gap suggests. In plain terms: you may pay a quarter of the wage and still get a comparable cost per unit if your output per worker is a quarter as high. The lever that closes that gap is not wages — it is UPPH. Every point of balance efficiency you recover is unit-labour-cost advantage you actually keep.

That is also why "we're cheaper than China" is a dangerous thing to rest on. The wage advantage erodes every year; the productivity you build from line balancing does not. For the headcount dimension — paying for operators who are not on the line at all — see contract labour headcount.

What you need to measure UPH and UPPH honestly

The trap is that UPH is easy to over-report. Count only good units at the end of the line, over the full paid time (including breaks if you want TEEP-style honesty), and divide by the operators actually assigned. The two inputs people fudge:

• Good vs total units. Counting units made rather than units that passed inflates UPH and hides the rework loop. If you count manually, read manual vs automated production counting first.
• Operator count. UPPH collapses if you quietly exclude helpers, loaders and the rework bench. Count everyone whose hours the line consumes.

Where a camera fits — and where it does not

Line balancing needs one thing a stopwatch gives you only once: the cycle time of every station, all shift. Nobody can stand and time eight benches for eight hours, every day. A camera pointed down the line does exactly that — timestamps each station's cycle, surfaces the bottleneck, and shows which neighbours are starved and which pile up WIP. That is the raw material of a rebalance, measured continuously instead of sampled once. See production line monitoring with cameras.

Be honest about the limit. A camera reads time and motion — cycle, idle, queue, balance. It does not read quality or test results — a solder joint, a crimp, a functional-test pass. Those stay with AOI, SPI and your jigs. And it is a process record, not a people score: we time the station, we do not rank the operator.

That is what Mama does with a short video of your line — return the station-by-station timing, the bottleneck, where hours are lost, a camera plan and a proposal. On your existing cameras or ours. For the full three-zone view of where an EMS line loses time, see where a mid-size EMS line loses hours.

FAQ

What is a good UPPH? There is no universal number — it depends entirely on your product's work content. UPPH is useful as a trend against your own line, not a benchmark to copy from another plant. Baseline it this month, rebalance, and beat last month.

How is UPPH different from UPH? UPH is total line output per hour; UPPH divides that by the number of direct operators, giving output per person. UPH can rise while UPPH falls if you add people — which is why UPPH, the labour-productivity number, is the one that tracks cost per unit.

How do I find my bottleneck without expensive software? Time every station over a representative hour. The station with the longest cycle is the bottleneck; the line runs at its speed. Doing this by hand across all stations at once is the hard part — a camera timing each station continuously removes the guesswork.

Does rebalancing really not cost anything? Moving work between existing stations costs only the industrial-engineering effort to do it — no new hardware, no new hires. The constraint is knowing the real per-station times, which is a measurement problem, not a capex one.

Can cameras calculate UPH automatically? They can time stations and count unit flow where there is a clear visual cue, which gives you a continuous UPH and the per-station data for balancing. They cannot confirm a unit is good — that is a test/inspection result — so pair camera timing with your quality data for a true good-UPH.