My Yield Is 3% Below the Recipe: Where Does the Product Go? (Indian Food Plants, ₹)
Yield 3% below the recipe: where did the product go? In an assumed 10 t/day bakery-snack plant, a 3-point gap is 300 kg of material a day: about ₹4.7 lakh a month at an assumed ₹60/kg blended cost (imported RBD palmolein alone quoted ₹1,44,000/tonne by SEA India on 16 Oct 2026). It sits in seven buckets, and you can weigh six.
By the Mama Editorial Team. Last checked: 17 Oct 2026, against the MoSPI press note on ASI 2024-25 (30 Sep 2026), the MoFPI Annual Report 2024-25, the SEA India quotes of 16 Oct 2026 and USDA FoodData Central. Every ₹ figure in the worked example is a labelled assumption.
Key takeaways
- A yield gap is a mass-balance problem, not a mystery. Weighed material in must equal packed product out, plus everything you threw, purged, swept, evaporated or gave away. What does not balance is the real leak.
- Do the balance on a dry-matter basis. Water leaves in the oven or fryer and oil goes in. Wet-weight yields hide both.
- Seven buckets cover almost every gap: moisture beyond recipe, start-up and changeover purge, spillage and sweepings, rejects, giveaway, unrecorded rework, and unexplained loss (counting errors, short receipts, pilferage).
- Three points of material is big money in food. Per ASI data, food products' value added is under 8% of their output, so 3% of material is roughly a third of what the plant adds.
- Cameras cannot weigh or measure moisture. They can show when the line purged, spilled, ran rejects or moved crates, which tells you where to put the scale.
What "3% below the recipe" actually means
The recipe, or standard yield, says how many kilograms of packed product a kilogram of input should give. If the standard says 94 kg of packed product per 100 kg of ingredients and the month closes at 91 kg, you are 3 points short. That is 300 kg a day on a 10-tonne day.
Be careful which 3% you mean. "Three percent below" can mean 3 points of input (94 to 91) or 3% of the standard (94 to 91.2). Agree on one definition with accounts first; this article uses points of input.
The gap also depends on what you call "input". If input is purchase invoices rather than weighed issues to the line, short supplier deliveries and store losses show up as a production yield problem. The first job is to weigh what is issued to the line, every shift.
Why the gap matters more in food than elsewhere
Food is the largest employer in Indian registered manufacturing. Per the MoSPI press note on ASI 2024-25 (released 30 Sep 2026), food products engaged about 23.6 lakh people, more than any other industry (MoSPI, ASI 2024-25 press note).
Food is also a thin value-added industry. In ASI 2022-23, as reported in the MoFPI Annual Report 2024-25 (Table 2.2), food products held 14.02% of registered manufacturing output but did not make the top five industries by gross value added, where the fifth-placed industry held 7.34% (MoFPI Annual Report 2024-25).
Our arithmetic on those two shares and the all-India totals in the same table: food's value added was below 8% of its output in 2022-23 (an upper bound of 7.94%). Inputs were therefore more than 92% of output.
In ASI 2024-25 food products did enter the top five by GVA, per the MoSPI press note of 30 Sep 2026, but that note gives no food GVA-to-output ratio (checked 17 Oct 2026).
When inputs are over 92% of output, losing 3% of input costs about 2.8% of output. That is more than a third of the value the industry adds. In a thin-margin business, the yield gap is often the profit gap.
The national post-harvest loss study described in the MoFPI Annual Report 2024-25 (NABCONS, submitted August 2022) measured losses from harvesting through storage and transport. Factory processing losses are not among its listed stages (checked 17 Oct 2026). No published Indian benchmark for in-plant snack or bakery yield was found in that report either. Your own scale is the benchmark.
Step one: balance the line on dry matter
Water is the biggest number in most food yields, and it moves in both directions. Dough gains water at the mixer and loses it in the oven. Potatoes lose most of theirs in the fryer and pick up oil.
A worked illustration from USDA FoodData Central, the US government food composition database: raw potatoes are 79.25% water, and plain salted potato chips are 1.86% water and 34.0% fat (FDC IDs 170026 and 169677, USDA FoodData Central).
Our arithmetic on those figures: one tonne of potatoes carries about 207.5 kg of dry solids. If every gram of solids reached the pack, it would make about 323 kg of chips. That is the ceiling before peel, slicing loss or a single broken chip.
Bakery works the same way in reverse. All-purpose wheat flour is 11.9% water and white bread is 36.4% water, per USDA FoodData Central (FDC IDs 168894 and 174924). A loaf weighs more than its flour because water was added and only part of it baked off.
So run the balance on dry solids, using your own lab moisture on input and output. Then a 1-point over-bake shows up as a moisture loss you can see, not as missing product.
The seven buckets, and where each is weighed
| Bucket | Where it happens | How to measure it (kg per shift) | Who owns the number | Can a camera see it? |
|---|---|---|---|---|
| Moisture beyond recipe | Oven, fryer, dryer | Lab moisture on input and packed product; compare to recipe target | QA | No |
| Start-up and changeover purge | First minutes after start, flavour or SKU change | Weigh every purge bin before it leaves the line | Shift in-charge | Timing and duration, yes; weight, no |
| Spillage and sweepings | Mixer discharge, conveyors, transfer points, seasoning drum | Weigh sweepings at shift end, in a separate bin | Line supervisor | Spill events and their timing, often yes |
| Rejects (broken, burnt, off-shape, metal-detector) | Cooling, sorting, packer infeed | Weigh reject crates; separate by reason | QA / production | Crates filling and leaving, yes; reason, partly |
| Giveaway | Filler / weigher | Checkweigher average minus declared net weight × packs | Packing in-charge | No |
| Unrecorded rework | Broken product fed back into dough or re-seasoned | Weigh rework issued back to the line and log it | Production planning | Re-feed events, often yes |
| Unexplained | Stores, receipts, WIP counts, gate | What is left after the six above | Owner / accounts | Crate and bag movements at doors, partly |
Giveaway has its own maths, covered in Product Giveaway in Indian Food Packing. You can price your own number with the product giveaway calculator.
Rework is the bucket that fools the most plants. Broken biscuit fed back into the next dough is not lost, but if nobody weighs it, the same kilogram is counted as input twice. Your yield then looks worse than it is, and the real loss hides behind it.
The worked example: a 10 t/day plant, in rupees
Assume a bakery-snack plant that issues 10 tonnes of dry ingredients a day (flour, sugar, fat, seasoning) and runs 26 days a month. The recipe standard is 94% packed output on ingredients, and the month closed at 91%.
Assume a blended raw-material cost of ₹60/kg. For scale, the fat in that blend is the dearest item: the Solvent Extractors' Association of India quoted imported RBD palmolein at ₹1,44,000 per tonne on 16 Oct 2026, about ₹144/kg (SEA India quotes).
The split of the 300 kg below is illustrative; weighing gives you your own.
| Bucket (assumed split) | kg/day | ₹/day at ₹60/kg | ₹/month (26 days) |
|---|---|---|---|
| Moisture beyond recipe (over-bake) | 60 | 3,600 | 93,600 |
| Start-up and changeover purge | 70 | 4,200 | 1,09,200 |
| Rejects not weighed or logged | 60 | 3,600 | 93,600 |
| Giveaway at the filler | 40 | 2,400 | 62,400 |
| Spillage and sweepings | 30 | 1,800 | 46,800 |
| Unexplained (counts, receipts, rework, gate) | 40 | 2,400 | 62,400 |
| Total gap (3 points of 10 t) | 300 | 18,000 | 4,68,000 |
At material cost, the gap is ₹4.68 lakh a month, or about ₹54 lakh over an assumed 300 working days.
Valued at selling price, the number is larger. Assume a net ex-factory realisation of ₹120/kg and a plant that sells everything it makes in season. Then the same 300 kg is ₹36,000 a day of sales, about ₹9.4 lakh a month.
If you can sell every packet, price the gap at selling price. If demand limits you, price it at material cost.
How to run the balance in two weeks
Week one: get the scales in place. Put a platform scale at the issue point to the line and one near the packing hall. Give purge, sweepings, rejects and rework each a coloured bin, and weigh every bin before it is emptied.
Week one: write it on one sheet per shift. Material issued, packed output (cartons × net weight, plus checkweigher average), and the four weighed bins. Ask QA for one moisture reading on input and on packed product per shift.
Week two: close the balance daily. Packed + purge + sweepings + rejects + moisture loss + giveaway should equal issued material. The remainder is the unexplained bucket. If it is under half a point, your yield problem is in the named buckets and is fixable on the line.
Week two: look at the shape, not the average. Purge that spikes on days with four SKU changes is a scheduling problem. Rejects that climb in the last two hours of a night shift point to the oven or fryer drifting. Sweepings that double on one line point to a specific transfer point.
If the unexplained bucket stays above a point, look upstream and at the doors. Re-weigh a sample of incoming bags against the invoice. Count WIP at shift change, not only at month end. The gate and weighbridge checks in The Gate and the Weighbridge apply here.
What cameras can and cannot tell you here
A camera cannot weigh anything, read a moisture meter or tell you a pack is 3 g over. Those numbers come from your scales, checkweigher and lab.
What video can do is put a time on events that drive the weighed buckets. It shows when the line started and when good product first reached the packer, so purge minutes can be compared with purge kilograms. It shows reject crates filling and leaving, spills at a conveyor, and rework being tipped back into a mixer.
At doors and the dispatch bay, video shows crates and bags moving, which helps reconcile the unexplained bucket. Keep that to material and zones. Counting crates is fine; following named workers is not what a yield balance needs.
For the time side of the same line, see how cleaning windows are treated in Sanitation-Adjusted OEE for Indian Food Lines. For the wider cost map of one product family, see Where a Biscuit and Confectionery Plant Loses Money.
Where Mama fits
Mama reads the video from a factory's existing cameras, or from cameras added where coverage is missing. Each morning it sends the owner a short, plain-language note on where hours and money leaked the day before. On a food line that means how long start-ups and changeovers ran before good product reached the packer, when reject and rework crates moved, and how long the line stood. It watches machines, zones and material flow, not individual people; the kilograms still come from your scales. To see which of the seven buckets your current cameras could put a time on, leave an email or phone in the form below, and Mama will get back to you with a camera plan.
FAQ
What is a normal yield loss for a snack or bakery line in India?
There is no published Indian benchmark for in-plant snack or bakery yield in the MoFPI Annual Report 2024-25 (checked 17 Oct 2026). The useful benchmark is your own recipe standard, balanced on dry matter, tracked shift by shift.
Why does my yield look worse in the monsoon?
Moisture in incoming flour, potatoes or grain changes with the season, so the same wet weight carries different dry solids. If the recipe standard is set on wet weight, it drifts with the weather. Balancing on dry matter, with a moisture reading on input every shift, removes that effect.
Is product giveaway part of the yield gap?
Yes. Overfill leaves the plant inside packs, so it never shows up as waste, but it is still product you paid for and did not sell. It is often a smaller bucket, but the easiest to measure, from the checkweigher log.
Should rework count as loss?
Rework fed back into the line is not lost, but it has to be weighed and logged as re-issued material. If it is not, the same kilogram is counted twice as input, and your yield looks lower than it really is.
Can cameras measure yield loss directly?
No. Cameras cannot weigh product or measure moisture. They can show when purges, spills, reject runs and crate movements happened, which tells you where to put a scale and which shift to look at first.