The number nobody can produce
Ask a moulding shop owner what their material wastage is and you will usually get a percentage, delivered with confidence. Ask how it was calculated and the confidence drops. In most cases it is a figure someone arrived at years ago, possibly from a supplier, possibly from a peer, and it has been repeated since without ever being measured against the shop’s own weighbridge.
The reason it is never measured is not laziness. It is that the arithmetic looks like it should already work. You know the part weight. You know the shot weight. You know how many shots the machine ran. Multiply and you have consumption. That model is clean, it is what the costing sheet uses, and it is wrong in a specific and expensive way: it only counts plastic that became a good part.
Consider a worked example. A shop runs a 40 g part on a four-cavity cold-runner tool. Shot weight is 160 g of part plus a 38 g runner and sprue — 198 g of PP through the nozzle for 160 g of saleable output. Before a single rejection, before a single colour change, the tool is running at roughly 81% material yield. The costing sheet, working from part weight alone, believes it is running at 100%.
Now add the rest of a normal shift. Two masterbatch changes, each needing the barrel purged. Fifteen startup shots after each change, discarded. A 3% rejection rate at inspection on parts that already consumed their full shot. Runners collected in a bin that gets ground on Thursday and tipped back into the hopper on Friday by a supervisor who does not tell anyone. At the end of the month, stores has issued materially more than the costing model says was consumed, and the difference becomes a stock adjustment nobody investigates because investigating it would require data that was never captured.
If your stock adjustments on polymer are consistently in one direction, that is not a counting error. It is your wastage, arriving as an accounting entry instead of a production report. The size of the recurring adjustment is a reasonable first estimate of what you are not tracking.
The six places granule actually goes
Wastage is not one number. It is six different mechanisms with six different fixes, and lumping them into a single percentage is what makes the problem feel unsolvable. Separate them and most shops find that two of the six account for the bulk of the loss, and one of those two is usually cheap to fix.
1. Sprue and runner
On a cold-runner tool this is structural, predictable, and the largest single component in most shops. It is also the one you can calculate rather than measure: weigh a complete shot, weigh the parts from it, and the difference is your runner weight per cycle. It does not vary. That makes it the easiest number to put into a BOM and the one whose absence distorts costing most.
2. Purging on grade and colour change
Every changeover pushes material through the barrel that will never be a part. The quantity depends on barrel capacity, the severity of the colour change — going from black to natural is not the same as natural to black — and operator patience. This is the component most sensitive to scheduling rather than engineering: sequencing a shift light-to-dark and grouping same-grade jobs can cut it substantially without changing anything on the machine.
3. Startup and stabilisation shots
The shots between pressing start and the process holding dimension. Predictable per changeover, and therefore proportional to how many times you change over. A shop running long campaigns barely notices it; a job-work shop doing six short runs a shift is paying it six times.
4. Rejections
The expensive category, because a rejected part has consumed its full shot weight, its full cycle time and its full machine cost before being rejected. A short shot found at the machine costs material. The same defect found at final inspection after a secondary operation costs material plus everything downstream. Where the rejection is caught matters as much as the rate.
5. Regrind that was never booked
Runners and rejects get ground and returned to the hopper without a stock transaction. Physically nothing is lost — and that is exactly why it is dangerous. The material reappears in production without reappearing in the system, so consumption looks lower than it is on the day the regrind is used and higher than it is on the day the virgin was issued. It corrupts the trend for both.
6. Moisture, spillage and handling
Hygroscopic grades such as nylon and PC lose weight in drying, bags get torn, hoppers get emptied between grades and the residue swept out. Individually small, collectively real, and unlike the other five this one is genuinely hard to eliminate. It is best treated as a known tolerance rather than a target — but you can only call it a tolerance once the other five are measured, otherwise it becomes the excuse that absorbs all of them.
Weight-based costing, built in
OEMup costs moulded parts by shot weight, runner weight and regrind blend ratio rather than by piece — so yield per part and per machine is a report, not a month-end reconstruction.
Book a 20-minute demo →Regrind is inventory, not free material
The most common costing error we see on a moulding floor is treating regrind as though it costs nothing. The logic is intuitive — it is material you already own, sitting in a bin, and using it feels like recovering something otherwise lost. But regrind carries the full acquisition cost of the virgin granule it came from, plus grinding, plus the handling to blend it. Costing it at zero does not make a job cheaper; it moves the cost onto whichever job consumed the virgin material originally, and quietly makes high-regrind jobs look more profitable than they are.
There is a quality dimension that matters more. Polymer chains degrade with each heat history. How many times a given material has been through the screw is a real process variable, and if regrind is untracked, it is an invisible process variable. Defects that correlate with regrind ratio then look random, because the one number that would explain them was never recorded.
Treating regrind properly is not complicated:
- Make it a distinct stock item — not the same item code as virgin.
PP-H110-REGRINDis a different material fromPP-H110, because it behaves differently. - Record the blend ratio in the BOM, so a part specified at 30% regrind is costed and planned at 30% regrind, and a deviation is visible.
- Book the grinding as a transaction — scrap out, regrind in, with weights. This single discipline closes most of the phantom-consumption problem on its own.
- Cap and track reuse generations where the grade and application warrant it, so material does not cycle indefinitely.
Contaminated material, purge lumps and mixed-colour scrap are usually worth more sold than reused, and reusing them imports a quality risk into a good job. Sold scrap is revenue with a clean tax treatment; blended contamination is a rejection you have not had yet.
The material balance that actually closes
Everything above reduces to one operational practice. Once per shift, per machine, weigh four things and record them against the job:
Shift material balance — worked example
That is the whole system. Four weights and a subtraction. It fits on a clipboard, it takes an operator a few minutes at handover, and it converts an argument into a number. Note what it gives you that a shot-weight model cannot: saleable yield of 76% on this shift, of which the runner is 17.5 points — recoverable through regrind — while purge, startup and rejection together are 5.5 points that are genuinely gone.
The unaccounted line is the important one. If it sits under about a percent, your measurement is trustworthy and you can start optimising the categories above it. If it is large or erratic, you have a data-capture problem, not a wastage problem, and no amount of process improvement will show up until the balance closes. Fix the measurement first.
What a spreadsheet tracks vs what a moulding shop needs
| Requirement | Typical Excel / Tally setup | What the shop actually needs |
|---|---|---|
| Material consumption | Part weight × quantity produced | Shot weight × shots, plus runner, purge, startup and rejection, reconciled to issue |
| Regrind | Not tracked; treated as free | Distinct stock item with batch identity, blend ratio in the BOM and a booked grinding transaction |
| Part costing | One material rate per part | Virgin and regrind at their own rates, by grade and masterbatch, with the runner allocated |
| Yield | Derived at month end, if at all | Per shift, per machine, per part — visible next morning |
| Scrap | Sold informally, recorded as miscellaneous income | Inventory item with HSN, weighed, invoiced, e-way billed where applicable |
| Material at job-work moulder | Invisible to planning once it leaves | Your stock at their location, netted by MRP, reconciled for ITC-04 |
| Grade and colour changeover cost | Absorbed into overhead | Attributed to the job that caused it, so scheduling can be optimised against it |
The India-specific layer
Three things make this a different problem in an Indian moulding shop than in a textbook.
Scrap sale is a taxable supply. Plastic waste, parings and scrap fall under HSN heading 3915, and selling it is a sale like any other — invoice, applicable rate, and an e-way bill for movement above the state threshold. Rates on scrap categories have been revised more than once in recent years, so confirm the current rate with your accountant rather than trusting what was configured in your billing software two years ago. The practical consequence is that informal scrap disposal is not simply untidy bookkeeping; it is unbilled taxable supply, and it also means the recovery never shows up against the job that generated the scrap.
Job-work moulding keeps the material on your books. If you send granule to another moulder, that stock is still yours, sitting at their premises. It moves on a delivery challan, the quantities have to reconcile for the ITC-04 return, and the process loss at the job worker needs to be an agreed allowance recorded up front — not a discrepancy discovered at year end when neither side remembers the run. The failure mode we see most often is planning blindness: material at the job worker is invisible to MRP, so it gets bought again. Our guide to job work and ITC-04 covers the compliance mechanics in full.
Material is bought by the kilo and sold by the piece. That unit mismatch is where most Indian SME costing breaks, and it is not specific to plastics — the same structural problem shows up in auto-component shops working to an OEM schedule and in fabrication units buying steel by weight. What is specific to moulding is that the conversion factor is not stable: it changes with the tool, the cavity count, the regrind ratio and the grade. A single per-piece material rate cannot represent it.
Five mistakes worth avoiding
Mistake 1 — Benchmarking before measuring
Chasing an industry wastage percentage heard from a supplier or a peer, without knowing your own. Different tools, cavity counts, changeover frequencies and grades make cross-shop comparison close to meaningless.
Instead: measure your own figure for one month and treat that as the baseline. Your trend against yourself is the only benchmark that can drive a decision.Mistake 2 — Putting part weight in the BOM instead of shot weight
The BOM says 40 g because the part weighs 40 g. Every plan, every purchase requirement and every cost built on that BOM is then short by the runner, permanently and invisibly.
Instead: the BOM carries shot weight divided by cavity count, with the runner either allocated across parts or held as a distinct line that returns to regrind.Mistake 3 — Grinding without a transaction
Runners go into the grinder and come back to the hopper with no stock movement recorded. Consumption figures then swing for reasons nobody can reconstruct a week later.
Instead: scrap out, regrind in, both weighed. It is two entries per shift and it removes most phantom consumption on its own.Mistake 4 — Measuring wastage monthly
A monthly figure tells you something went wrong somewhere in twenty-six shifts across nine machines. It is a number you can report but not act on, and by the time it arrives the tool, grade and operator have all changed.
Instead: capture per shift and per machine. Aggregation upward is free; disaggregation after the fact is impossible.Mistake 5 — Treating startup and purge as unavoidable
They are unavoidable per changeover, but the number of changeovers is a scheduling decision, not a law of physics. Shops that sequence jobs by grade and by colour — light to dark — pay the penalty far fewer times.
Instead: attribute changeover material to the job that triggered it and let the planner see it. See capacity planning for small factories for how sequencing decisions get made in practice.Getting this running in about a month
This does not need a system before it needs a habit. The order below is what we use during onboarding, and it works on paper before it works in software.
Week 1 — Weigh a shot on every running tool
Complete shot, then the parts from it. The difference is your runner weight per cycle, per tool. This is a one-time exercise that immediately corrects every BOM you have, and it usually takes one person a single day.
Week 1 — Put a scale at the scrap bin
Nothing else changes yet. Runners, rejects and purge get weighed into separate bins at shift handover. You are establishing the measurement before you attempt to influence it.
Week 2 — Start the four-line balance
Virgin in, regrind in, saleable out, scrap out — per machine, per shift. Expect the unaccounted line to be large and erratic at first. That is normal, and it is a data-capture problem to fix before anything else.
Week 3 — Make regrind a real stock item
Separate item codes by grade, a booked grinding transaction, and a blend ratio recorded against the job. The balance should tighten noticeably within a few shifts of this landing.
Week 4 — Rebuild the BOMs on shot weight
Now that runner weights are measured and regrind has an identity, the BOM can carry shot weight, cavity count, grade and blend ratio. Costing and MRP both become correct at the same moment.
Ongoing — Yield per part, per machine, every morning
Once the balance closes reliably, yield stops being a monthly argument and becomes a daily report. That is the point at which improvement work has something to aim at.
Every step above works with a scale and a printed sheet. Software removes the transcription and makes the reporting automatic, but a shop that has not established the weighing habit will not get a trustworthy number out of any ERP, ours included.
When this is not worth doing
Some shops genuinely do not need most of this. If you run hot-runner tools on long single-grade campaigns with few changeovers and a stable rejection rate, your structural wastage is small and predictable, and the effort is better spent elsewhere. If you are a two-machine unit moulding one part for one customer, a monthly weighbridge reconciliation will tell you what you need to know.
The practice earns its keep when you have variety: multiple grades, frequent colour changes, cold-runner tools, meaningful regrind, or job work in either direction. Those are the conditions under which the six mechanisms diverge from each other, and a single percentage stops describing anything real.
The bottom line
Material is the largest line in a moulding shop’s cost sheet and usually the least instrumented. The gap is not caused by a shortage of technology — it is caused by a costing model built on part weight, which by construction cannot see runner, purge, startup, rejection or regrind. Replace that model with a four-line balance weighed per shift, give regrind an identity, and put shot weight into the BOM. The measurement is cheap, it works on paper, and it turns the largest uncontrolled cost on your floor into a number you can manage.
Everything after that — costing by grade, MRP that nets material at the job worker, batch traceability from granule to despatch — is ordinary plastics and moulding ERP, and it only works if the weights underneath it are real.
FAQ
What is an acceptable material wastage percentage in injection moulding?
There is no cross-shop number worth quoting. It depends on part geometry, runner design, changeover frequency and regrind policy. A hot-runner tool on a single grade wastes very little; a cold-runner job-work shop changing colour four times a shift loses a large fraction to runner and purge alone. Measure saleable kilograms divided by virgin kilograms issued, per part and per machine, and track your own trend.
Why doesn’t shot weight times shots match actual consumption?
Because shot weight counts only material that entered the cavity on a good cycle. It excludes sprue and runner, purge on changeover, startup shots, rejected parts that already consumed their shot, and moisture and handling loss. The difference gets written off as a stock adjustment.
Should regrind be treated as free material?
No. It carries the acquisition cost of the virgin material plus grinding, and it has a finite number of reuse cycles before properties degrade. Treating it as free understates cost on high-regrind jobs and makes regrind-correlated defects look random.
How is plastic scrap treated under GST?
Sale of plastic waste, parings and scrap is a taxable supply under HSN 3915, invoiced like any other sale, with an e-way bill for movement above the state threshold. Rates have been revised more than once — confirm the current rate with your accountant rather than relying on an old billing-software configuration.
How do you track material sent out for job-work moulding?
As your stock at someone else’s premises. It moves on a delivery challan, quantities reconcile for ITC-04, and process loss should be an agreed allowance recorded up front. The common failure is that material at the job worker is invisible to planning and gets purchased again.
Can a small shop do this without sensors or IoT?
Yes, and most should start there. A scale, a shift issue record and four weights per machine will close the balance to a workable tolerance. Automated capture improves resolution and removes transcription error, but it solves a measurement problem you should already have defined.
Related reading
- ERP for Plastics & Moulding Manufacturers — the product view: BOMs by mould and cavity, weight-based granule costing with regrind
- The Hidden Costs of Inventory — what unmeasured stock costs beyond the material itself
- Job Work and ITC-04 — the compliance mechanics when moulding goes out or comes in
- Multi-Level BOM: The Complete Guide — structuring BOMs that carry shot weight and blend ratio
- ERP for Auto Component Manufacturers — the same unit-mismatch problem under an OEM schedule
- Vendor Portal Software for Indian Manufacturers — keeping job-work partners inside the same stock ledger
- Production Cost Calculator — a quick way to test what a yield change is worth on your own numbers