Skip to content
allanninal.dev/build
Collaborate

Part 4 of 7 · Packaging data reporter series ~5 min read

Household or not, and why that field costs more than the rest of the file

Everything up to this point produced a tonnage, and a tonnage on its own does not cost anything. This one field decides which of those tonnes carry a disposal fee, which makes it the most expensive column in the file by a wide margin.

Key takeaways

  • Fees attach to the household portion. The non-household portion is reported and not charged.
  • It is a property of the sale, not of the product. The same SKU splits both ways.
  • The default is household, and the burden of showing otherwise sits with you.
  • Trade-only pack formats evidence themselves. Everything else needs the buyer.
  • Household tonnage is also split by nation, which the sales ledger already knows.

Classifying a sale, not a product

A sales line classified as household or non-household packagingA vertical chain inside an AWS account container, entered from a box on the left labelled A sales line with a product, a customer and a quantity. Sold to a consumer, directly or through a retailer, with a side exit reading yes: Household, this tonnage carries a fee. Trade-only format, meaning a pack no household would ever buy, with a side exit reading yes: Non-household, evidenced by the format itself. Evidence on the buyer, from customer type, channel or delivery address, with a side exit reading yes: Non-household, evidenced by who bought it. The final step is Household by default, because you cannot show otherwise. A note says the default costs money and that this is deliberate, because the burden of showing that packaging is not household sits with the organisation reporting it.AWS ACCOUNTA sales linea product, a customer,a quantitySold to a consumer?direct, or througha retailerHouseholdthis tonnagecarries a feeyesTrade-only format?a pack no householdwould ever buyNon-householdevidenced by theformat itselfyesEvidence on the buyer?customer type, channel,delivery addressNon-householdevidenced by whobought ityesHousehold by defaultbecause you cannotshow otherwiseThe default costs money, and that is deliberate. The burden of showing that packaging is not household sits with the organisation reporting it.
Fig 1. Three questions and a default that is not free. Every line ends up classified, and the reason it was classified that way is stored with it.
  • Machine learning
  • Management
  • Analytics
  • People

Where the 310 tonnes came from

In the worked example the wholesaler classified by product code, which is the obvious way to do it and requires one decision per line rather than one per sale. Every code that had ever been sold to a consumer was marked household, and the bulk catering packs of those same products inherited the marking.

Three hundred and ten tonnes of packaging that went out in six-kilogram catering packs to restaurants and schools was therefore reported as household. Nobody made a mistake; the model was one field too coarse. Classifying at the sales line instead costs nothing extra, because the sales ledger already knows who bought it.

Three piles, and the third is the argument

Tonnes grouped by whether the household classification is obviousThree vertical bars showing tonnes placed on the market grouped by how clear the classification is. Clearly household accounts for nine hundred and thirty tonnes. Clearly not household accounts for six hundred and forty tonnes. Depends who bought it accounts for three hundred and ten tonnes. A note says the third bar is sixteen per cent of the tonnage and essentially all of the uncertainty, and it is also the only bar a sales ledger query can resolve.0500100015002000~930Clearly household~640Clearly not~310Depends who bought itTonnes placed on the marketThe third bar is 16 per cent of the tonnage and essentially all of the uncertainty. It is also the only bar that a sales ledger query can resolve.
Fig 2. The same 1,880 tonnes, grouped by how obvious the classification is. Two of these piles need no work at all.

Do not over-correct

Having found three hundred and ten tonnes on the wrong side, the instinct is to go looking for more, and that instinct needs a limit. An unevidenced non-household claim is a different kind of error from an over-cautious household one: the first is found in an audit and reversed with consequences, the second only costs money.

So the rule the system enforces is that non-household requires a stored reason and a stored piece of evidence, and household needs neither. That asymmetry mirrors where the burden actually sits, and it means the submission can be defended line by line rather than argued for in general terms.

Shipment packaging is household more often than people expect

The box an online order arrives in, the void fill inside it and the tape around it are packaging, they are supplied to whoever receives the parcel, and when that is a consumer at a home address they are household packaging. For a business with a direct-to-consumer channel this is often the fastest-growing line in the whole submission and the one least likely to be in the spreadsheet.

It also has the useful property of being entirely within your control, unlike the primary pack that came from a supplier. A business that discovers its shipping cartons are a hundred and eighty tonnes a year has a purchasing decision in front of it with a number attached to it, which is more than most compliance exercises produce.

What moves a tonne to non-household

  • Customer type on the ledger. Trade account, with a company number.
  • Channel. A wholesale or foodservice order route, coded at entry.
  • Pack format. A catering size or a trade multipack, evidenced by the spec.
  • Delivery address type. A business address, where you hold it.
  • A contract. Supply agreements that state the end use.
  • Nothing. In which case it stays household, and that is the correct answer.

The nation split underneath

Household tonnage is also reported by nation, because the disposal it pays for is organised nation by nation. This sounds like an extra dimension of work and is usually not, because the sales ledger already carries a delivery postcode and the mapping is a lookup.

Where it does bite is with a distributor whose customer is a national retailer’s central warehouse. The goods go to one address in one nation and are then sold in four, and the reporter has to allocate on the best information available. The system’s job there is to record the allocation basis rather than to pretend the postcode answered the question.

The next post is what actually gets submitted, what has to survive behind it for seven years, and why the submission is the smaller half of the deliverable.

All posts