The answer to how to reduce manufacturing waste is to measure it before you fix it. Record what each run takes in, the good output it gives and every scrapped unit with its cause. Two figures, yield and scrap rate, then show where the money goes and which cause to fix first.
This guide covers the kinds of waste, the measures, and one week at a Lagos bakery worked in naira. It ends with the fixes and how to record scrap in ArmPOS.
What are the types of manufacturing waste?
The classic list has seven wastes, named by Toyota engineer Taiichi Ohno. The Asian Productivity Organization lists them as waiting, transporting, processing, inventory, motion, defects and overproduction. It adds newer ones, such as untapped human potential, energy and water.
| Waste | What it looks like | How to see it |
|---|---|---|
| Overproduction | Bread baked for orders that never came | Loaves unsold at the end of the day |
| Waiting | Dough waiting for the only oven | Minutes between mixing and baking |
| Transporting | Trays carried across the room to cool | Metres walked per batch |
| Processing | Slicing and bagging loaves a buyer wants whole | Steps no customer pays for |
| Inventory | Flour and yeast kept past their dates | Stock age and expired batches |
| Motion | Bending and reaching for tins and tools | Reaches and steps per batch |
| Defects | Burnt, collapsed or crushed loaves | Scrap rate by cause |
| Untapped skills | Fixes the bakers know and nobody asks for | Ideas raised and tried each month |
| Energy and water | A generator running with no batch in the oven | Litres of diesel per batch |
Defects are the easiest waste to measure. Each one also carries every cost spent on it before it failed, so it is usually the place to start.
How do you measure manufacturing waste?
Measure three things for every run: what went in, the good output, and what was scrapped, by cause. Four figures come from those counts.
| Measure | Formula | What it tells you |
|---|---|---|
| Yield | Good output ÷ expected output from the materials used × 100 | How much material became goods you can sell |
| Scrap rate | Units scrapped ÷ units started × 100 | How much was lost outright |
| First-pass yield | Units right the first time, with no rework ÷ units started × 100 | How much went right without fixing |
| Cost of scrap | Units scrapped × cost per unit at the stage they failed | What the loss is worth in money |
Expected output comes from your recipe or bill of materials: what one bag, sheet or roll should give. Without that standard, yield has nothing to be measured against.
Keep scrap and rework apart. A reworked unit still sells, so it isn't scrap, but it cost labour twice and lowers first-pass yield.
For example: one week of waste at a Lagos bakery
For example, a bakery in Lagos bakes six batches a day, six days a week, each from one 50 kg bag of flour. Its recipe should give 100 loaves a bag. The figures are illustrative.
A loaf costs NGN 900 to make and sells to distributors at NGN 1,300. This was the week's count.
| Line | Loaves | Share of scrap |
|---|---|---|
| Expected from 36 bags | 3,600 | |
| Good loaves | 3,312 | |
| Scrapped: a batch lost to a power cut mid-bake | 120 | 41.7% |
| Scrapped: over-proofed and collapsed | 84 | 29.2% |
| Scrapped: burnt | 48 | 16.7% |
| Scrapped: crushed in slicing or packing | 36 | 12.5% |
| Total scrapped | 288 | 100% |
Yield is 3,312 ÷ 3,600 = 92.0%, and the scrap rate is 288 ÷ 3,600 = 8.0%. The bakers also reshaped 90 pieces of dough before proofing, so first-pass yield was 3,222 ÷ 3,600 = 89.5%.
The lost loaves cost NGN 259,200 that week, or NGN 13,478,400 over a year. Each one cost NGN 900 and earns nothing, so it wipes out the NGN 400 profit of 2.25 good loaves.
Two causes, the power cut and over-proofing, make up 70.8% of the loss. They are where to start.
Normal waste and abnormal waste
Decide what normal waste is, and treat the rest as a problem to fix. IFRS for SMEs makes "abnormal amounts of wasted materials, labour or other production costs" an expense (paragraph 13.13, module on inventories).
If this bakery sets its normal waste at 2%, or 72 loaves a week, the other 216 are abnormal. That is NGN 194,400 a week that no price should carry, as manufacturing costs explains.
How to reduce manufacturing waste, step by step
Count the losses, fix the biggest causes, then stop the next ones upstream. Six steps cover it, and each one ends with a count.
1. Record every loss with its cause
Give each station a scrap tray and a tally sheet with four or five causes. Count the trays at the end of each shift, while a loss can still be traced to a batch, an oven or a person.
2. Rank the causes and fix the top two
Sort the week's scrap by cause, biggest first, and work on the top two until they shrink. Fixing a small cause first feels busy and saves little.
| Cause | Fix | Loaves lost a week: before | After | Saved a week (NGN) |
|---|---|---|---|---|
| Power cut mid-bake | Start no batch the generator can't finish: check its fuel before each one | 120 | 20 | 90,000 |
| Over-proofing | A timer, and a board showing when each batch started to prove | 84 | 24 | 54,000 |
| Burnt | Check the oven with a thermometer before the first batch | 48 | 18 | 27,000 |
| Crushed | Cooling racks at slicing height, and one tray per batch | 36 | 21 | 13,500 |
| Total | 288 | 83 | 184,500 |
The four fixes save 205 loaves and NGN 184,500 a week, or NGN 9,594,000 a year. Scrap falls from 8.0% to 2.3%, close to the 2% the bakery calls normal.
3. Stop defects upstream
Look behind each defect for a machine out of adjustment or a material past its date. Service ovens, mixers and slicers on a calendar, not when they fail.
Use flour, yeast and other materials earliest expiry first, as raw materials inventory sets out.
4. Make what will sell, not what the oven can hold
Overproduction becomes scrap when goods expire unsold. Bake to orders and to the same day's sales last week. Plan festive peaks such as Christmas and Eid weeks ahead, in a production plan.
The same holds for materials: buying far ahead of use is waste with a delay. Reorder points, set out in how to avoid overstocking and stockouts, keep stock to what each delivery needs to cover.
5. Clear the bottleneck, then cut the walking
Waiting comes from a bottleneck: the one station that sets the pace for all the others. At the bakery, one oven makes the dough wait, and dough that waits over-proofs.
Transporting and motion come from where things stand. A production layout that follows the steps cuts both.
6. Ask the people who scrape the trays
Ask the bakers which loss they would fix first, and try their answer for a week. The person who scrapes a burnt tray usually knows why it burnt.
Then count the trays again. A fix that doesn't move the numbers wasn't a fix.
How to record manufacturing waste in ArmPOS
ArmPOS has no production module: no bill of materials, no work orders and no automatic use of materials when goods are made. A small maker can still record each run's input, output and losses, and read its yield from the records.
- Record each run with two adjustments. A Subtraction takes off the materials used, and an Addition puts on the goods made. Add only the good output, and write the counts in the Note, such as "Run 9: 100 expected, 88 good, 12 burnt" (Count and adjust stock).
- Record goods lost after the run as damages. Damages › Create Damage takes broken or spoiled stock off, with a note on what happened. Use a damage or an adjustment for the same goods, never both (Record damaged or spoiled stock).
- Total the losses. The damages list exports to Excel, and the Stock Adjustment Report totals what each run added and removed, with each adjustment's cost (Stock reports).
- Track materials that expire by batch. Switch on Track batches & expiry, and the Expiry Report lists what expires within the warning window (Batches and expiry dates).
- Put each machine on a service date. In Assets, record the oven and the mixer with a Next Validation date. ArmPOS emails the people whose role includes Assets when one is due (Company assets, on some plans).
The run method is the one set out in barcodes in manufacturing. A damage books nothing in Accounting: if you use it, record the loss at cost with a manual journal, as inventory write-off shows.
Frequently asked questions
What are the 8 wastes of lean manufacturing?
Taiichi Ohno's seven: overproduction, waiting, transporting, processing, inventory, motion and defects. Lists of eight add untapped skills: what the people doing the work know and nobody uses.
How do you calculate scrap rate?
Divide the units scrapped by the units started, and multiply by 100. At the bakery: 288 ÷ 3,600 = 8.0%.
What is the difference between yield and first-pass yield?
Yield counts all good output, reworked or not. First-pass yield counts only what was right the first time: 92.0% and 89.5% at the bakery.
What is a good scrap rate?
Your own normal level, taken from your best recent weeks, then lowered step by step. A figure from another industry tells you little about a bakery or a furniture workshop.
Can all manufacturing waste be removed?
No. Some waste is normal, like offcuts and trimmings, and it stays in the product's cost. The aim is to remove the abnormal part and lower the normal one over time.
Put a tally sheet at each station this week and count the scrap every evening. To record runs, losses and expiry dates in one place, create a free ArmPOS account and add your first material as a product.
