An inventory can be entirely accurate and still be wrong. Every figure in it correct, every factor appropriate, every calculation reproducible — and the total materially understated, because a whole category of emissions was never recorded at all.
- Accuracy is what teams worry about. Completeness is what gets flagged in assurance.
- An empty category looks exactly like a category that genuinely has no emissions.
- Refrigerant leakage, waste, upstream transport and upstream fuel emissions are the usual omissions.
- Asserting immateriality without ever estimating it is a gap, not a judgement.
Accuracy is what teams worry about. Completeness is what gets flagged.
Why the two failures behave differently
A misfiled record is visible. It sits in the wrong category, and sooner or later somebody reviewing the breakdown notices that business travel is filed under Scope 1.
A category with nothing in it looks like a category with no emissions. There is no row to review, no figure to question, no anomaly in the chart. It reads as a clean zero, and a clean zero reads as a fact.
This is why an inventory can look complete and quietly not be. Nothing about the interface distinguishes “we measured this and it was nil” from “nobody thought about this”.
The ones most Southeast Asian manufacturers miss
From working through real inventories, the same four categories go missing again and again.
Refrigerant leakage (Scope 1, fugitive). A tropical manufacturing site with chillers is losing gas. It is not a question of whether — refrigeration systems leak, that is why they get recharged. The emissions are recorded on a contractor’s service sheet, in a filing cabinet, and nobody in the sustainability team has ever asked for a copy. Because the gases have warming potentials in the thousands, a modest recharge is a meaningful tonnage. This is the single most commonly missing line we see.
Waste (Scope 3, category 5). Every plant produces waste and pays somebody to take it away. That contractor sends invoices and waste transfer notes. The emissions depend heavily on treatment — landfill, incineration and recycling of the same tonnage give very different answers — so tonnage alone is not enough, which is often why it gets postponed and then forgotten.
Upstream transport (Scope 3, category 4). Raw materials arrive. Somebody paid for that freight. If the vehicles are not yours, it is not Scope 1 mobile combustion, and it is not in your electricity bill, so it belongs to a category that has no natural owner in the organisation and therefore no natural source of data.
Fuel- and energy-related activities (Scope 3, category 3). The upstream emissions of producing and delivering the fuel and electricity you bought. It is entirely derivable from figures you already have, which is precisely why it is skipped: no new data collection is required, so no data collection process surfaces it.
Notice the pattern. These are not missed because they are hard to measure. They are missed because nobody in the organisation currently owns them. Electricity has an owner in facilities. Fuel has an owner in fleet. Refrigerant, waste, inbound freight and upstream fuel emissions belong to nobody, and unowned data does not get collected.
Immaterial and unmeasured are not the same thing
The usual defence for an empty category is materiality. Some categories genuinely do not apply — a services business has no process emissions, and saying so is correct and complete.
The failure is asserting immateriality without ever having estimated it. “We assumed employee commuting was immaterial” is a judgement. “We have not looked at employee commuting” is a gap. They are frequently the same sentence in a first report, and an assurance provider will ask which one it is.
The distinction is cheap to fix and expensive to leave. A rough estimate — headcount times a typical commuting distance times a published factor — takes an afternoon and produces one of two useful outcomes. Either the number is small, and you have a documented basis for excluding it. Or it is not, and you have found something.
What you cannot defend is a blank with an assumption attached to it after the fact.
Where the assistant helps, and where it does not
Our platform runs a completeness check against the inventory. It reads what is recorded, what the organisation actually does, its facilities and countries, and reports categories that have nothing in them but almost certainly should.
On our demonstration manufacturer it produced five, and the reasoning is the useful part. Not “you are missing fugitive emissions” — anyone can produce that from the category list. Instead:
The boundary shows a manufacturing plant in Johor and a Singapore head office. The inventory has no fugitive emissions records despite likely refrigeration and air-conditioning at both sites. Only diesel and electricity are recorded across three records total, so refrigerant leakage is unreported.
To close it: preventive maintenance logs for HVAC and refrigeration, service invoices and top-up records, plus a site equipment inventory listing chillers, refrigerant type and charge.
That second paragraph is what separates a useful gap analysis from a checklist. “Do an audit” is not actionable. “Ask your HVAC contractor for twelve months of service invoices” is a task somebody can complete this week.
What it does not do is decide anything. Every gap lands as a proposal. Rejecting one requires a written reason, which is how “considered and not applicable” gets recorded rather than being something you have to reconstruct in the assurance meeting a year later.
What to do this quarter
If your inventory is in its first or second year, the highest-value hour you can spend is not improving a figure you already have.
Print the fifteen Scope 3 categories. Go through them with someone from operations and someone from procurement, not from sustainability, because the sustainability team already knows what it knows. For each one, write down one of three things: the number, who would have the data, or a specific reason it does not apply to you.
You will end with a short list of owners to approach and a shorter list of documented exclusions. Both of those are worth more to your next report than another decimal place on Scope 2.