Material Retention Rate · Working Paper v1 Manufacturing · Supply Chain · Circularity
A field note from the factory floor

We counted where the waste went.
Nobody counted what we kept.

Zero-waste-to-landfill measures a destination. Recycled content % measures an inflow. Neither one asks the question a mill manager has asked for a hundred years: of every kilogram that entered this plant, how much is still material?

That question has a name. Material Retention Rate.

01In the product
shipped mass
02Retained at grade
same alloy, polymer, fibre
03Downcycled
sold cheap, grade lost
04Dissipated
burned, buried, oxidised

Illustrative split — the four fates of every kilogram bought. MRR counts 01 + 02. Diversion metrics count 01 + 02 + 03 + energy recovery.

Three candidate north stars — only one measures retention

The metric decides the behaviour.
Two of these reward the leak.

Candidate A

Zero Waste to Landfill

Certifies the destination of waste, not its existence. Under common diversion schemes, tonnage sent to energy-from-waste can still count toward the claim, and a tonne of high-grade aluminium sold as low-grade deoxidiser counts as fully "recycled". A plant can double its scrap generation and improve its diversion score.

Measures exit route

Candidate B

Recycled content %

Measures what came in, and says nothing about the 10–50% of that material that leaves the same building as offal, sprue, skeleton or chip. Worse: under ISO 14021, material reworked inside the process that generated it cannot be claimed as recycled content at all — so the factory that keeps its own scrap earns nothing, while the factory that ships it out lets someone else book it as pre-consumer content.

Measures inflow only

Candidate C

Material Retention Rate

Measures the share of purchased mass that is still in productive material form at the same grade — inside the product, or returned to an equivalent use. It is yield thinking, extended past the factory gate to the whole tier-1 and tier-2 chain. It cannot be improved by making more scrap or by burning better.

Measures retention

Why the current pace will not arrive

The circular share of the world economy is going backwards while extraction climbs.

6.9%

of all material entering the global economy is secondary — down from 9.1% in 2018 and 7.2% the year before.

Circularity Gap Report 2024 ↗
106Bt

billion tonnes of virgin material dug, pumped, felled and harvested each year to feed it.

Circle Economy ↗
60%

projected growth in global resource extraction by 2060 on the current trajectory, says UNEP.

Global Resources Outlook 2024 ↗
26×

US non-hazardous industrial solid waste (~7.6 bn tons/yr) versus municipal solid waste (~292 Mt/yr). The mass is upstream, in the factory — not in the household bin.

US EPA — Sustainable Materials Management ↗

Consumer-facing circularity has been improving faster than the system it sits in. Across the Ellen MacArthur Foundation's Global Commitment signatories, recycled content in plastic packaging roughly doubled from 4.8% in 2018 to around 11–12% — genuine, hard-won progress on one side of the ledger. On the other side, the converters, moulders, mills and cut-and-sew floors making that packaging still generate double-digit process scrap that no brand report has a line for.

Definition — back to the raw question

MRR, stated plainly

Retention is not diversion. A kilogram is retained only if it can do the job it was bought for — same alloy family, same polymer grade, same fibre length, same purity class. So the numerator is grade-weighted, and the denominator is everything you paid for.

MRR = ( Mproduct + Σ ( qi × Mreturned,i ) ) ÷ Mpurchased

q = 1.00  returned to the same grade and the same application (closed loop)
q = 0.75  returned to the same material family, lower spec (open loop, in-sector)
q = 0.40  downcycled — grade destroyed, mass survives
q = 0.00  incinerated, landfilled, oxidised, dissipated, or unknown

Three boundaries, reported separately. One number for the whole chain hides where the loss lives. Report them as a stack, the way a mill reports home, prompt and obsolete scrap.

MRR-1 · PROCESS

Machine to machine. Coil in, part out. Sits next to first-pass yield and OEE on the same shift board, owned by the same production manager.

MRR-2 · SITE

Gate to gate. All purchased mass across the site, including packaging, consumables, solvents, purge and start-up rejects.

MRR-3 · CHAIN

Tier-1 and tier-2. The converter, the mill, the dye house, the cell plant. Where most of the loss actually is — and where nobody currently reports.

The scrap ledger — where the mass goes before anyone sees a product

Loss is designed in,
long before disposal.

Reported industry ranges, gathered from process literature and company disclosures. They are order-of-magnitude anchors for a first conversation — the only numbers that matter are the ones from your own bill of material and your own weighbridge.

40–50%

Aluminium sheet stamping (body-in-white)

Blanking and trimming send up to half the coil out as offal and skeleton before a panel exists.

Lever · alloy-segregated offal, baled at the press, contracted back to the same caster
80–95%

Titanium aerospace machining

Buy-to-fly ratios of 10:1 to 20:1 mean most of a certified billet leaves as chips and coolant-wet swarf.

Lever · near-net-shape deposition, dry/segregated chip handling, re-melt qualification
5–30%

Li-ion cell manufacturing

Electrode coating, calendering and formation reject 5–10% at steady state, and 20–30% during a ramp.

Lever · in-plant black mass loop; cathode metals returned to the same cell line
10–30%

Injection moulding & extrusion

Sprue, runner, purge and start-up shots. Regrind is technically easy; contamination and colour discipline are not.

Lever · hot runners, colour-segregated regrind bins, capped regrind ratios by spec
10–20%

Garment cutting rooms

Marker efficiency decides it. Mixed-fibre, mixed-colour "jhut" is then sold by the sack at near-zero value.

Lever · sort at the cutting table by fibre + colour; traceable pre-consumer feedstock
20–30%

PCB panel utilisation

Rails, frames, break-outs and test coupons — laminate, copper and gold leaving as board offcut.

Lever · panelisation design review; segregated recovery of precious metal fractions
Who is already doing this — with numbers

Six loops that are closed
at the machine, not at the bin.

Jaguar Land Rover × Novelis
Automotive aluminium · closed loop

The REALCAR programme is the textbook MRR intervention: press-shop offal is segregated by alloy at the press, never commingled, and contracted straight back to the caster that supplied the coil. It returns as RC5754 — the same grade, for the same panels. Recycled aluminium takes about 95% less energy than primary metal, so retention here is an energy and a cost decision before it is an environmental one.

Same-grade returnTens of thousands of tonnes/yr reclaimed~95% energy saving vs primary

Sources: Novelis · JLR Sustainability · International Aluminium Institute

Apple
Consumer electronics · supplier-side scrap recovery

Apple's 100% recycled aluminium enclosure alloy was built substantially on its own manufacturing chips — the machining scrap from unibody CNC operations, kept clean and returned to the smelter rather than sold into the open market. That is MRR-3 in practice: the brand reached into a supplier's process and changed how the scrap is handled. Reported: 22% of shipped material recycled or renewable, 56% recycled cobalt in batteries, and a public target of 100% recycled cobalt.

100% recycled Al enclosures56% recycled cobalt22% recycled/renewable shipped mass

Source: Apple Environmental Progress Report

Redwood Materials × cell makers
Batteries · the shortest loop in industry

The largest battery "recycler" in North America is, in mass terms, largely a manufacturing scrap business: production reject from gigafactories, processed and returned as cathode and anode material to the same plants. Recovery rates of over 95% for nickel, cobalt, copper and lithium mean the metal never leaves the industry. Tesla similarly reports recovering thousands of tonnes of nickel, copper and cobalt from its own manufacturing streams.

>95% recovery Ni · Co · Cu · LiFeedstock = factory scrap

Sources: Redwood Materials · Tesla Impact Report

Aquafil — ECONYL
Polymers · depolymerisation back to virgin grade

Nylon 6 taken back to caprolactam monomer and repolymerised, so pre-consumer yarn waste and carpet fluff return at virgin specification rather than as a downgraded filler. Aquafil's published figure: every 10,000 tonnes of ECONYL avoids around 70,000 barrels of crude oil and 65,100 tonnes of CO₂-equivalent. This is q = 1.00 retention on a polymer that would otherwise be lost.

Virgin-grade regeneration70,000 bbl crude avoided / 10 kt

Source: Aquafil ECONYL · Interface (ReEntry take-back)

Norsk Titanium
Aerospace · designing the loss out

The cheapest scrap is the scrap never cut. Rapid Plasma Deposition builds a near-net-shape preform and cuts buy-to-fly from roughly 10:1 to about 2:1 — meaning the material saving comes from process design, not from a better bin. MRR rewards this; a diversion certificate is completely blind to it.

Buy-to-fly 10:1 → ~2:1Loss designed out, not diverted

Source: Norsk Titanium

Reverse Resources · Bangladesh
Textiles · making pre-consumer waste traceable

Cutting-room waste has always had a market — it is just an invisible, informal, quality-destroying one. By digitally tracing each factory's waste by fibre composition and colour, segregated at source, the same tonnage moves from low-value shoddy into fibre-to-fibre recyclers at several times the price. The material didn't change. The information did.

Segregate at the cutting tableTraced by fibre + colour

Source: Reverse Resources · Textile Exchange standards

And the oldest case of all. Steelmakers have measured home scrap, prompt scrap and obsolete scrap separately for over a century, because yield loss was a cost line before it was ever an ESG line. MRR is not a new invention. It is the mill's oldest instinct, taken out of the mill and applied to the chain.

Sources: worldsteel · IEA Iron & Steel Technology Roadmap

The organisational fault line
Keep it as pure as it came — just in more pieces.
The one sentence version

In most companies, waste is one team and recycled content is another. They talk. They do not work as one. Waste management is measured on cost per tonne removed; materials sourcing is measured on price per tonne bought. Nobody is measured on the mass that survives between them — so the offal is commingled for collection convenience, the grade is destroyed in the skip, and the same company then pays a premium to buy recycled material back from the open market.

Starting from waste management, you will never arrive. Start instead from the recycled-content suppliers you already qualify: go and look at how they handle their own material, how they trace it, how much they scrap, and whether that scrap comes back pure. One owner, one mass balance, one number.

The audit — questions to ask, in order

Six stages. Ask them of your own plant first,
then of every tier-1.

The sequence matters: you cannot segregate what you have not weighed, cannot trace what you have not segregated, and cannot contract for what you cannot trace. Run it in this order and it is an engineering project. Run it out of order and it becomes another waste report.

STAGE 1 — MASS BALANCE

Do we know the kilograms?

  • How many tonnes of each material entered this site last month, by grade and by supplier?
  • How many tonnes left as product, and how many as everything else?
  • Do the two sides reconcile to within 2%? If not, where is the unexplained mass?
  • Is scrap weighed at the machine, or estimated once at the skip?
STAGE 2 — SEGREGATION AT SOURCE

Where does the grade die?

  • At which exact operation do two different grades first touch each other?
  • Are offcuts binned by alloy / polymer / fibre and colour at the machine, or after the sweep?
  • What contaminates the stream — coolant, adhesive, label, coating, mixed fastener?
  • What would it cost to add one more bin, and what is that grade worth intact?
STAGE 3 — TRACEABILITY

Can we prove the identity?

  • Does each scrap batch carry the material spec, lot and origin that its parent coil carried?
  • Who is the second buyer of our scrap, and the third? Do we know where it lands?
  • Can we evidence the loop for a Digital Product Passport or a recycled-content claim?
  • Is the data in the ERP, or in a broker's paper ticket?
STAGE 4 — RETENTION BY DESIGN

Can the loss be removed?

  • What is the theoretical minimum input mass for this part — the buy-to-fly, the marker efficiency, the nesting yield?
  • Which three design decisions are creating the most scrap, and who signed them off?
  • What regrind or re-melt ratio does the spec actually permit — and who set that limit, on what evidence?
  • Does the product spec forbid retention for a reason we can still defend?
STAGE 5 — COMMERCIAL TERMS

Who owns the kilogram?

  • Who legally owns process scrap under our current supply contract — us, the supplier, or the hauler?
  • Is there a scrap-back / toll-return clause returning segregated material to the original mill?
  • Is the price a flat per-tonne removal fee, or indexed to the grade retained?
  • Does anyone in procurement have MRR in their objectives, or only unit price?
STAGE 6 — GOVERNANCE

Who is accountable?

  • Is there one named owner for MRR across waste, operations and sourcing — or three partial ones?
  • Does MRR appear on the plant board next to OEE and first-pass yield, in the same font?
  • Do we report MRR alongside recycled content %, so improvement can't come from making more scrap?
  • What is the target trajectory — and what happens to the bonus if it is missed?
Ask these before the next scrap tender is signed
Implementation — first year

Not a new programme.
A new line on the existing one.

First 90 days

Weighbridge and machine-level mass balance for the top three materials by spend. No new systems — reconcile purchase records against shipped mass and skip tickets. Publish the gap, however embarrassing. Name one owner.

By month six

Segregation at source on the two largest loss streams. Rewrite the waste contract from removal-fee to grade-indexed. Add scrap-back clauses to the next three tier-1 renewals. Audit two recycled-content suppliers on their own MRR.

By month twelve

Grade-weighted MRR-1, MRR-2 and MRR-3 reported quarterly next to recycled content %, mapped onto WBCSD CTI circular-inflow/outflow and GRI 306 disclosures. Design reviews carry a yield gate. The metric now moves money.

This paper is not published because the work is finished. It is published because an old, fixed mindset cannot keep up with a fast-growing problem — and because most of the answer is already sitting in the source code of how we make things. Go and read it.

References & sources

Everything above, traceable.

  1. Circle Economy — Circularity Gap Report 2024. Global circular material use rate; annual material extraction.
    circularity-gap.world/2024Global circularity data
  2. UNEP / International Resource Panel — Global Resources Outlook 2024.
    unep.org — Global Resources Outlook 2024Extraction trajectory to 2060
  3. US EPA — Sustainable Materials Management & industrial materials recycling.
    epa.gov/smmIndustrial vs municipal waste mass
  4. Ellen MacArthur Foundation — Global Commitment progress reporting.
    ellenmacarthurfoundation.org/global-commitmentRecycled content in packaging
  5. WBCSD — Circular Transition Indicators (CTI).
    ctitool.com% circular inflow / outflow framework
  6. GRI 306: Waste 2020 disclosure standard.
    globalreporting.org/standardsWaste disclosure boundaries
  7. ISO 14021 — self-declared environmental claims; pre-consumer / post-consumer definitions.
    iso.org — ISO 14021Why in-process rework can't be claimed
  8. UL Solutions — Zero Waste to Landfill validation (UL 2799).
    ul.comDiversion claim methodology
  9. GBCI — TRUE zero waste certification.
    true.gbci.orgDiversion certification scheme
  10. Textile Exchange — Recycled Claim Standard / Global Recycled Standard.
    textileexchange.org/standardsPre-consumer input rules
  11. Regulation (EU) 2023/1542 on batteries and waste batteries — recycled content and recovery efficiency targets.
    eur-lex.europa.eu — EU Battery RegulationMandated recycled content
  12. European Commission — circular economy policy, ESPR & Digital Product Passport.
    environment.ec.europa.eu — circular economyTraceability regulation
  13. Novelis — closed-loop aluminium recycling.
    novelis.comPress-shop scrap loops
  14. Jaguar Land Rover — sustainability & REALCAR closed loop.
    jaguarlandrover.com/sustainabilityAlloy-segregated offal return
  15. International Aluminium Institute — recycling energy data.
    international-aluminium.org~95% energy saving vs primary
  16. Apple — Environmental Progress Report.
    apple.com/environmentRecycled aluminium, cobalt data
  17. Redwood Materials — battery manufacturing scrap recovery.
    redwoodmaterials.com>95% metal recovery
  18. Tesla — Impact Report.
    tesla.com/impactIn-house scrap metal recovery
  19. Aquafil — ECONYL regenerated nylon.
    econyl.comDepolymerisation to virgin grade
  20. Interface — ReEntry take-back and materials programmes.
    interface.comProduct take-back loops
  21. Norsk Titanium — Rapid Plasma Deposition, buy-to-fly reduction.
    norsktitanium.comNear-net-shape yield
  22. worldsteel — steel scrap and recycling fact sheets.
    worldsteel.orgHome / prompt / obsolete scrap
  23. IEA — Iron and Steel Technology Roadmap.
    iea.org — Iron & Steel RoadmapScrap availability and yield
  24. Reverse Resources — traceability of pre-consumer textile waste.
    reverseresources.netCutting-room waste value chain

Figures are quoted as reported by the sources above and as commonly cited industry ranges. Verify every range against your own bill of material before using it in a target.