Original illustration of two complete packaging systems protecting equivalent fragile bottles, one paperboard and one reusable crate; neither is presented as lower carbon.
Carbon Footprint of Packaging Materials: Compare the Whole Pack, Not Just the Material
Diagram text description

Original synthetic editorial illustration. The featured artwork depicts two plausible complete packaging systems carrying the same fragile product, without asserting either has lower emissions. Spatial grouping, branches, proportions or directional paths carry the claim with concise labels.

A paper wrap can have a lower production factor per kilogram than a plastic wrap and still produce a worse result for a particular shipment. It may need more mass, an added moisture barrier, or a larger outer carton. If it protects the product less effectively, damaged or spoiled goods can erase the apparent saving. The reverse is also possible. The carbon footprint of packaging materials becomes a useful comparison only after the alternatives are made equivalent in the job they perform.

Begin with one question: how can we deliver the same specified quantity of product in acceptable condition, through the same route and use period, with the lowest defensible greenhouse-gas impact? Compare the complete packaging systems needed to do that job. Count their components and realistic disposal paths; where protection differs, include the resulting product loss. Record where data are measured, estimated or missing. A material name or a generic factor is an input, not the answer.

This article is a decision framework for a packaging change, not a list of universally best materials. The European Commission's packaging and plastics lifecycle methods use the function of the protected product and the required packaging flow as the comparison basis. That matters more than a headline such as “paper beats plastic” or “recycled always wins.”

The answer depends on what the packaging must protect

Packaging does not exist to provide kilograms of material. It keeps a product usable while it is filled, stored, transported, displayed and opened. The service can be easy to overlook in a footprint spreadsheet because a material factor is readily available while damage or shelf-life performance takes work to establish. A fair comparison starts with the service, then calculates the quantity and type of packaging each option needs to provide it.

Consider a beverage delivered in a carton and in a bottle. Comparing one kilogram of carton board with one kilogram of bottle polymer does not tell you how many saleable litres each package delivers. Nor does it capture barrier layers, caps, refrigeration requirements, breakage or the distribution route. A glass bottle might be reused; if so, a defensible comparison needs the actual collection and wash system and a plausible number of trips. A single-use alternative must be assessed as such. The result can change with those conditions without either material being inherently “good” or “bad.”

The JRC's plastics lifecycle method illustrates a functional unit built around delivering a specified volume of beverage without breakage and with a defined shelf life. That is not a claim that every packaging project must use the same beverage example. Its transferable lesson is to state what, how much, how well, and for how long the pack must deliver before applying emission factors. If the two proposed packs do not deliver the same service, do not publish their footprints as a like-for-like material ranking.

The packaging system includes the layers around the product.
The packaging system includes the layers around the product.
Diagram text description

Read from the outside inward: outer shipper, cushioning, primary pack and product. These are nested physical components, not quantities to assume for every design. Closures, labels and liners belong in the corresponding bill of materials even where the schematic does not draw each separately.

Define one packaging job for both options

Write a one-sentence job specification before requesting supplier data: “Deliver 1,000 saleable units of this product to the named market, on this distribution route, with the required barrier, shelf life and damage tolerance.” Then define the reporting period and the point at which a unit counts as successfully delivered. The wording turns a vague material comparison into a testable procurement decision.

Table 1: scroll horizontally if needed.

Decision field Keep constant or explicitly test
Product and quantity Same item, fill volume or weight, and number of saleable units
Performance Protection from moisture, oxygen, light, contamination and breakage as applicable
Route and period Same warehousing, shipping distance, handling and required shelf life
Outcome Same acceptable condition at receipt, with losses recorded separately

Some fields cannot be held constant because a new package changes them. That is precisely why they belong in the model. A lighter pack may lower freight emissions; a weaker seal may increase returns. Run those as linked scenario differences rather than quietly assuming identical loss rates. If the team lacks reliable loss data, show a sensitivity range: how much additional spoilage would overturn the apparent packaging saving? Do not claim that the threshold has been crossed until measured experience or a defensible trial supports it.

Agree on the job specification with packaging engineering, quality and distribution teams. They know which performance failures matter. The sustainability team can then map those requirements into a functional unit and an evidence request rather than trying to infer them from a catalogue description.

Build a bill of materials for 1,000 successful deliveries

Count every component needed to do the packaging job. A headline material name can conceal a liner, seal, label, cushioning or additional outer box. Record the mass of each component per packed unit, then calculate the reference flow required for the chosen number of successful deliveries. Use measured samples or supplier specifications with a clear product version.

Consider two fictional designs for the same product, initially assuming equal protection and no losses. Design A uses 20 g of polymer and 5 g of label or closure per unit. Design B uses 30 g of board, a 4 g liner and a 3 g label or closure. The example is an inventory calculation; it does not establish which pack has lower emissions.

Table 2: scroll horizontally if needed.

Component Design A per unit Design B per unit Quantity for 1,000 units
Main material 20 g polymer 30 g board A: 20 kg; B: 30 kg
Liner None in this example 4 g A: 0 kg; B: 4 kg
Label or closure 5 g 3 g A: 5 kg; B: 3 kg
Complete pack 25 g 37 g A: 25 kg; B: 37 kg

The board design therefore requires 12 kg more packaging mass for this assumed delivery count. That is not a carbon ranking: the materials, production routes and other life-cycle stages differ. Its purpose is to prevent the 4 kg liner and 3 kg closure from disappearing behind a “board packaging” label.

Convert grams to kilograms before applying a factor expressed per kilogram. Check whether a supplier's per-pack factor already includes conversion, scrap or transport so those stages are not added again. If protection differs, the number of packs required to deliver 1,000 successful units also changes. The next example introduces that effect rather than assuming the initial 1,000-pack inventory remains sufficient.

Complete pack mass differs from the mass of its headline material.
Complete pack mass differs from the mass of its headline material.
Diagram text description

For the initial equal-protection 1,000-pack inventory, Design A contains 20 kg polymer and 5 kg closure or label, totalling 25 kg. Design B contains 30 kg board, 4 kg liner and 3 kg closure or label, totalling 37 kg. Bars show mass, not greenhouse-gas impacts.

Use factors that match the supplier and production route

A generic “paper” or “plastic” factor can hide large differences in raw-material route, recycled share, manufacturing energy and location. Ask suppliers whether their number represents the actual grade and production plant, a product category average or a secondary database. Record the factor's geographic and temporal coverage, included processes and allocation approach. An apparently precise comparison can be dominated by incompatible assumptions rather than a real design difference.

A recycled-material factor can vary with the dataset and allocation method. Before comparing two suppliers, align the production boundary and identify which processes their values include. A factor for resin or board production alone is not equivalent to a converted, printed and transported pack. Where a supplier cannot disaggregate its number, use a consistent secondary-data approach for screening and identify the missing supplier detail as a priority for the next design decision.

Recycled content requires special care. Verify which component contains it, the percentage by mass, the evidence for the claim and the factor's treatment of recycling. “Recyclable” describes a technical possibility; “recycled content” describes an input; actual local collection and processing determine an end-of-life pathway. They are not interchangeable fields. Where two suppliers report different factor boundaries, ask for a harmonized calculation or show the difference as an uncertainty, not as a conclusive win for one material.

Calculate the product-loss threshold before choosing a lighter pack

This is a separate fictional impact scenario from the earlier mass inventory. Its assumed impacts are not calculated from those material quantities.

The relevant output is a successfully delivered product, not simply a dispatched package. If a new design changes breakage or spoilage, compare the complete systems needed to achieve that output. The following fictional example uses deliberately simple assumptions to show the effect; its values are not measured packaging or product footprints.

Assume each produced item has an impact of 2.00 kg CO2e before packaging. Pack A adds 0.10 kg per item and has a 1% loss rate before successful delivery. Pack B adds 0.06 kg per item and has a 3% loss rate. Assume failed items are replaced, every dispatched item uses a pack, and the loss rate remains constant. Other stages are held equal and omitted from this teaching comparison.

To deliver 1,000 successful items, expected production is 1,000 divided by the success rate. Pack A requires 1,000 ÷ 0.99 = 1,010.10 items and packs. Pack B requires 1,000 ÷ 0.97 = 1,030.93. Fractional counts are expected averages across repeated shipments, not instructions to order part of an item.

Table 3: scroll horizontally if needed.

System Expected items and packs Impact per dispatched item Total for 1,000 successes
A 1,010.10 2.00 + 0.10 = 2.10 kg CO2e 2,121.21 kg CO2e
B 1,030.93 2.00 + 0.06 = 2.06 kg CO2e 2,123.71 kg CO2e

Despite its lower packaging impact, B is about 2.50 kg CO2e higher in this defined scenario. The difference is small, so uncertainty could easily change the ranking. The example is useful because it identifies the deciding measurement: protective performance.

For B to break even, solve 2.06 ÷ (1 − loss B) = 2.10 ÷ 0.99. The resulting loss threshold is approximately 2.886%. Below that threshold B is lower under these assumptions; above it A is lower. This is a percentage of dispatched items, not an additional percentage on top of A's loss rate. A real trial should measure the relevant failure mode and include changed distribution, replacement and waste-treatment impacts where material. Use the threshold to design that trial, not as proof that either actual package will win.

Different loss rates change how many products and packs are needed for the same successful delivery count.
Different loss rates change how many products and packs are needed for the same successful delivery count.
Diagram text description

With the fictional 1% loss rate, A needs 1,000 divided by 0.99, or 1,010.10 expected dispatches. With 3% loss, B needs 1,000 divided by 0.97, or 1,030.93. Fractional quantities are expectations across repeated shipments. Each branch represents an alternative way to provide the same delivery service.

Include distribution, reuse and end-of-life

The carbon footprint of packaging materials does not stop at manufacturing. Record where components are made, converted, filled, shipped and collected. A heavier package may alter transport emissions; a flat-packed format may improve vehicle utilization; an aseptic format may change storage requirements. These effects are context-dependent. Do not simply multiply material mass by a freight factor and assume the same route, load factor and handling for both designs.

Reuse needs a system, not a symbol.

Count the reusable container's production, collection, sorting, washing, losses and replacement across a plausible number of completed uses. If a return rate is unknown, show the break-even number of successful turns and test a lower-turn case. A theoretical ten-use design used only twice should not be reported as if it delivered ten uses. The same logic applies to pallet, crate and refill systems, including reverse logistics.

Count completed uses, including the work required to return a container to service.
Count completed uses, including the work required to return a container to service.
Diagram text description

The container moves through use, return and sorting, washing, and next use. The loop is a real service cycle; not every container completes it. Lost or rejected containers leave the loop, requiring replacement and the relevant treatment scenario. Count completed turns rather than theoretical design life.

At end of life, distinguish what a material can do in principle from what happens in the target market. A composite may need specialized collection; a paper or plastic item can be contaminated by the packaged product; metal and glass have different local recovery routes. The model should state the specific collection, sorting, recycling, incineration or landfill scenario and the basis for choosing it. Do not assign recycling benefits twice—once as recycled input and again as an unexplained disposal credit. If allocation conventions differ across data sources, reconcile them before presenting a comparative number.

Completed uses are the denominator for the reusable container: production divided by two successful uses gives five times the production allocation per use compared with ten, before washing and return transport are added. Count failed returns as losses from the loop. Once a container leaves service, route it through the local treatment scenario instead of continuing to credit hypothetical future trips.

The route out of service depends on local collection and treatment.
The route out of service depends on local collection and treatment.
Diagram text description

A pack leaving service can enter a local recovery pathway or another treatment pathway. Select the actual scenario, such as recycling, incineration or landfill, with its allocation rules. The split is qualitative and gives no assumed percentages or automatic recycling credit.

Read the comparison as a decision table

Combine the bill of materials, protective performance and life-cycle assumptions in one comparison. The decision table brings the separate evidence requirements together; complete it using one consistent dataset for the real project.

Table 4: scroll horizontally if needed.

Decision input Evidence to obtain What it can change
Complete pack Weighed components and supplier specification Required material quantities
Protection Comparable damage, spoilage or returns trial Number of items needed per successful delivery
Production Matched supplier or process factors Impact of each component
Distribution Actual routes, loads and handling Outward and replacement transport
Reuse Completed trips, return and wash data Production allocated across successful uses
End of life Local collection and treatment scenario Disposal and recycling treatment

Begin with the rows that could reverse the decision. In the fictional loss calculation, a trial distinguishing 2% from 3% loss is more useful than adding precision to the assumed packaging factor. In a reusable system, actual completed turns may dominate instead. The important uncertainty depends on the design and service, so the same checklist should not produce the same priority for every project.

Report the result per successful delivered unit and identify the main contributors. If the preferred option changes under a plausible alternative loss rate or recovery pathway, show both scenarios. That finding can justify a targeted pilot rather than an immediate material switch. Where the evidence supports a stable result, record the conditions under which the choice holds so later supplier or route changes can be assessed against them.

What the result can and cannot claim

A matched comparison can support a conditional statement: for the specified product, route, performance, period and lifecycle boundary, the proposed pack's modeled greenhouse-gas result is lower or higher under the stated assumptions. It cannot, without broader evidence, support “paper is always lower than plastic,” “recyclable means low carbon,” or “our product is carbon neutral.”

A carbon-only comparison also cannot establish that the package has the lowest overall environmental impact; water use, resource use and litter are different questions.

Before any external claim, review whether the result is sensitive to uncertain factor choices, product-loss rates or reuse turns. Check whether a public comparative claim requires an independent critical review under the applicable method and market rules. The public ISO 14044 abstract expressly includes interpretation, reporting, critical review and limitations in lifecycle assessment; the exact obligations depend on the study and intended disclosure. Do not imply that an internal option screen is automatically a verified product declaration.

If the model's strongest evidence is a generic database factor and no one has tested actual protective performance, say “preliminary scenario estimate” and keep the conclusion internal. The decision may still be valuable: it can justify a supplier data request or a pilot design. The wrong move is to turn a provisional calculation into a confident material label. The published claim should be narrower than—or at most equal to—the evidence.

Run a packaging trial around the deciding uncertainty

Turn the sensitivity result into a practical test with packaging engineering and quality teams. Define the product, route, handling conditions and failure criterion before dispatch. Record the number of units sent and received successfully, the type of failure and whether the package or another cause was responsible. Compare designs under sufficiently similar conditions to make the result useful.

For the fictional loss example, the question is whether B's failure rate is reliably below the approximately 2.886% break-even threshold under the model's assumptions. A small trial with no observed failures does not establish a zero underlying failure rate. The trial size and uncertainty need appropriate statistical treatment, especially when the decision depends on a narrow difference. The article's arithmetic does not prescribe a universal sample size.

Use the trial to collect secondary effects as well: extra cushioning, damaged-product replacements, changed pallet loading or additional handling. Update the same comparison rather than treating the prototype's laboratory performance as the whole distribution story. If the route or product changes, decide whether the test still represents the intended use.

The immediate next step is a jointly specified trial and one complete comparison file. Keep the measured quantities separate from scenario assumptions and identify which result would change the design decision. This gives procurement a specific evidence request and a clear basis for choosing the next design.

In the fictional model, B only wins below the break-even loss rate.
In the fictional model, B only wins below the break-even loss rate.
Diagram text description

Under the article’s fixed fictional impact assumptions, B is lower only when its loss rate is below approximately 2.886%. At equality the model ties; above it B is higher. Measurement uncertainty must be considered before deciding which branch applies. This is not a universal packaging threshold.

Conclusion: optimise the delivered product, not the material label

Compare the packaging systems required to deliver the same product successfully. The bill-of-materials example shows why a liner or closure matters; the loss calculation shows why a cheaper carbon result per pack can disappear when more products must be replaced. Neither example establishes a universal preference for paper, plastic or another material.

Start with a common delivery job, count the complete packs and identify the uncertainty most likely to change the result. Then gather the supplier data or run the protective-performance trial that resolves it. A decision supported by those quantities and conditions is more useful than a broad material slogan, and easier to revisit when the product, route or package changes.