The inventory is built. Every category is mapped, the supplier data is tagged, the factors are sourced, the estimates are labelled with their uncertainty. Now comes the moment that decides whether all of that work survives: the carbon accountant has to declare the inventory closed and hand it to an external assurer who will test it against a threshold. Not every number gets the same scrutiny. The assurer is looking for misstatement that matters, and the whole question of closing is whether the inventory is complete and accurate enough, where the categories that move the total are solid and the gaps that remain are disclosed rather than hidden. Close it wrong and you have spent months building something that fails the engagement. Close it right and the same file that satisfies the assurer is the one that proves the work. This lesson is about closing a Scope 3 inventory that survives.

What Closing Actually Means

Closing is not the moment you stop working; it is the moment you assert that the inventory is fit to be assured, and that assertion is itself a claim the assurer tests. To close defensibly you have to know what the assurer is testing against, because they are not checking every number to the last decimal. They are testing whether the inventory is free of material misstatement: an error or omission large enough that it could change a reasonable reader's understanding of the company's emissions. A tiny error in a small category does not threaten the inventory; a large error or a silent omission in a category that drives the total does. Closing means getting the material categories solid and the rest honestly disclosed, then asserting that the whole is fit for assurance.

This reframes the entire endgame. You do not chase perfection across all fifteen Scope 3 categories, the value-chain emissions that average roughly 75% of a footprint. You concentrate effort where misstatement would matter, and you disclose, rather than hide, the limitations of the rest. The skill of closing is allocating your remaining time and your remaining data-collection effort to the categories and lines where being wrong would actually change the conclusion, and accepting honest, disclosed roughness everywhere it would not.

The Materiality Threshold the Assurer Tests

The materiality threshold is the line below which a misstatement is judged too small to matter to the reader's understanding, and above which it must be corrected. In financial audit this is often a quantified figure; in sustainability assurance it is applied with judgment, but the logic is the same: the assurer sets a sense of what magnitude of error in the emissions total would be material, and tests whether the inventory could contain an error that large undetected. For Scope 3, this immediately concentrates attention, because the categories are wildly unequal in size. A company whose footprint is dominated by Category 1 purchased goods and services and Category 11 use of sold products will see the assurer spend almost all their effort there, because those are the only categories where an error could be material.

Understanding the threshold lets you close intelligently. Identify your material categories, the handful that together make up the bulk of the total, and make those solid: high primary-data coverage where achievable, well-sourced factors, defensible labelled estimates with disclosed uncertainty for the parts you cannot measure. For the small categories, far below the threshold, a coarse spend-based estimate, clearly labelled, is perfectly defensible, because even a large proportional error in a tiny category cannot move the total enough to matter. The mistake teams make is spreading effort evenly, polishing a tiny category while a material one rests on thin data. Close against the threshold and you put your effort exactly where the assurer will put theirs.

You are not closing a perfect inventory. You are closing one where the categories that move the total are solid and the gaps that remain are disclosed, not hidden. Completeness you can prove beats precision you cannot.

Coverage: The First Thing the Assurer Checks

Coverage is the share of your emissions, and your material categories, that is actually accounted for rather than omitted. It is the first thing an assurer assesses, because an inventory can be beautifully accurate on what it includes and still be materially misstated by what it leaves out. A company that reports Scope 3 but quietly excludes a material category has not produced a smaller inventory; it has produced a misstated one. So closing requires a coverage check: have all material categories been included, and is every exclusion documented with its basis? An excluded category that is genuinely immaterial, documented as such, is fine. An excluded category that is material, or excluded with no stated reason, is the classic finding.

Coverage also has a within-category dimension. Inside a material category, what share of the activity is covered by primary data, what share by labelled estimates, and what share, if any, is simply missing? An assurer wants to see that the category is complete, even if much of it is estimated, rather than complete-looking because the gaps were averaged away invisibly. The honest position is a category that says: here is what we measured, here is what we estimated and how, and here is the small remainder we could not get and have disclosed. That is higher coverage in the sense that matters, because nothing is hidden, even though not everything is measured.

Completeness Versus Precision: The Trade-Off You Manage

Closing forces a trade-off that teams often get backwards. Completeness is whether all material emissions are accounted for; precision is how exactly each included figure is known. Faced with limited time, you have to choose where to spend it, and the assurance logic is clear: completeness beats precision. A complete inventory with some honestly-disclosed coarse estimates is more defensible than a precise-looking inventory that silently omits a material category. An assurer can work with a rough but disclosed estimate; they cannot work with a hidden hole, because the hole is exactly the kind of omission the materiality threshold exists to catch.

This does not mean precision is irrelevant. Within a material category, precision matters where imprecision could push the figure across the materiality threshold. The discipline is to ask, for any given line: would a more precise figure change whether the inventory is materially correct? If yes, invest in precision there. If no, a disclosed estimate is enough. You manage completeness and precision as a single budget, spending precision effort only where it could change the conclusion, and spending the rest of your effort on making sure nothing material is omitted. Completeness you can prove is the foundation; precision is the targeted upgrade where it actually matters.

Limited Versus Reasonable Assurance: What Each Demands of Coverage

The assurance level changes what closing has to deliver. Limited assurance, the most common engagement today, has the assurer perform procedures sufficient to conclude that nothing has come to their attention suggesting the figures are materially misstated. It is a negative form of conclusion and a lighter touch: the assurer samples, asks questions, and reviews, but does not test exhaustively. Even so, coverage of material categories and documented exclusions are squarely in scope, because a missing material category is exactly what a limited-assurance review is designed to notice.

Reasonable assurance, the higher level the market is trending toward, has the assurer gather enough evidence to give a positive conclusion that the figures are fairly stated. This demands deeper, broader testing: more lines traced to source, more estimates reproduced, more coverage verified directly rather than reviewed. For closing, the move from limited to reasonable raises the bar on every dimension at once: coverage must be more demonstrably complete, estimates more thoroughly supported, and the whole inventory more deeply reconstructable from evidence. A company anticipating the shift to reasonable assurance closes its inventory to the higher standard now, so the transition is a confirmation rather than a scramble. In both cases the closing logic is the same: solid where it is material, disclosed where it is not, and reconstructable throughout.

The Basis-of-Preparation: What You Hand Over When You Close

Closing is not just an internal decision; it produces a document the assurer reads first, the basis-of-preparation. This is the methodological statement that explains how the inventory was built: the organizational and operational boundary, which of the fifteen categories are included and which excluded and why, the methods used for each category and the data sources behind them, the treatment of estimates and their uncertainty, and the basis for the materiality judgment that decided where effort went. The basis-of-preparation is where the closing decisions are made explicit and defensible, and a strong one does most of the work of the engagement, because it answers the assurer's structural questions before they are asked.

A weak or absent basis-of-preparation is itself a finding, regardless of how good the underlying numbers are, because the assurer cannot assess an inventory whose methodology is undocumented. The discipline at closing is to write the basis-of-preparation as the inventory is finalised, not after the assurer requests it, so that it reflects what was actually done rather than a tidy after-the-fact narrative. The closing assertion, that the inventory is fit to be assured, is made through this document: it is where you state, in writing, that the material categories are solid, the immaterial ones are estimated and labelled, the exclusions are justified, and the whole is reconstructable. When the document says these things and the inventory backs them up, closing is clean.

Defending the Closing Decision Under Challenge

Expect the closing decision to be challenged, and build it to withstand the challenge. The assurer may push on three fronts. They may question a materiality judgment, asking why a particular category was treated as immaterial, and your answer is an estimate or proxy showing its plausible magnitude sits far below the threshold, recorded as the basis for the call. They may question coverage, asking whether a material category is genuinely complete, and your answer is the measured-versus-estimated split showing nothing was averaged away invisibly. They may question an estimate, asking how uncertain it is and whether the uncertainty could approach the threshold, and your answer is the disclosed confidence band and, where the band could matter, the better data you invested in to tighten it. In each case the defence was built before the question, which is what separates a closing that survives from one that scrambles.

A Worked Example: Closing a Footprint Against the Threshold

A manufacturer is closing its Scope 3 inventory. The total comes to roughly 100,000 tonnes CO2e, and the categories are deeply unequal, which is typical.

The shape of the footprint. Category 1 purchased goods and services is about 62,000 tonnes, roughly 62% of the total. Category 11 use of sold products is about 28,000 tonnes, roughly 28%. Category 4 upstream transport is about 5,000 tonnes. The remaining categories, business travel, commuting, waste, and the rest, together come to about 5,000 tonnes, none individually above 2% of the total. The assurer's materiality threshold, applied with judgment, lands such that an error of a few percent of the total would be material.

Where the effort goes. The accountant concentrates almost everything on Categories 1 and 11, which together are 90% of the footprint. In Category 1, she drives primary-data coverage as high as she can get it, sources every material factor, and for the unmeasured remainder builds labelled spend-based estimates with disclosed uncertainty. In Category 11, she does the same for the use-phase model that dominates it. For Category 4 and the small categories, she uses clearly-labelled coarse estimates, because even a 30% error in a category that is 5% of the total moves the inventory by 1.5%, comfortably below the threshold, and chasing precision there would waste effort the material categories need.

The closing assertion. She documents coverage: all fifteen categories considered, the material ones built solid, the immaterial ones estimated and labelled, every exclusion documented with its basis (for example, no franchises, so Category 14 excluded with a stated reason). She records, for each material category, the split of measured versus estimated and the uncertainty on the estimated portion. She can now make the closing assertion honestly: the inventory is complete across material categories, accurate where accuracy matters, and transparent about its estimates and its limits. When the assurer tests it, they find their effort and hers landed in the same place, on Categories 1 and 11, and the file answers their questions there.

Compare the inventory that fails. It reports a clean 100,000-tonne total with every category looking equally finished, no visible split between measured and estimated, and Category 14 silently dropped with no note because the company assumed it was obvious. The total might be the same. But when the assurer asks about coverage, the file cannot show which categories are solid and which are thin, cannot show the measured-versus-estimated split in Category 1, and cannot explain the missing category. The same number, closed without regard to the threshold, becomes a finding instead of a clean opinion.

Notice what the disciplined close did with its scarce effort. It did not try to make Category 4 or the small categories precise, because no amount of precision there could change whether the inventory is materially correct. It poured that saved effort into Categories 1 and 11, where a few percentage points of error genuinely could cross the threshold, and into the basis-of-preparation that documents the whole. The result is an inventory that is faster to close, because effort was not wasted on immaterial precision, and more defensible, because the material categories are solid and everything is documented. The closing decision is, in the end, an allocation decision: spend your remaining time where being wrong would matter, disclose honestly everywhere it would not, and write down why you drew the line where you did. Done that way, the same file that closes the inventory quickly is the file that survives the engagement, which is the whole point of closing against the threshold rather than against an imaginary standard of total precision.

Key Takeaways

• Closing is the moment you assert the inventory is fit to be assured, and the assurer tests that assertion against material misstatement: an error or omission large enough to change a reasonable reader's understanding of the emissions.
• The materiality threshold concentrates the assurer's attention on the few categories that drive the total, so close by making material categories solid and disclosing, rather than hiding, the roughness of the rest.
• Coverage is the first thing the assurer checks: all material categories included, every exclusion documented with its basis, and within each material category an honest split of measured, estimated, and disclosed-missing.
• An excluded material category, or any exclusion with no stated reason, is the classic assurance finding; a documented exclusion of a genuinely immaterial category is fine.
• Completeness beats precision: a complete inventory with honestly-disclosed coarse estimates is more defensible than a precise-looking one that silently omits a material category, because the hole is what the threshold exists to catch.
• Manage completeness and precision as one budget: invest precision only on lines where a more exact figure could change whether the inventory is materially correct, and spend the rest ensuring nothing material is omitted.
• Limited assurance gives a negative conclusion from a lighter, sampled review, while reasonable assurance demands deeper, broader testing; closing to the reasonable standard now makes the transition a confirmation rather than a scramble.
• In the worked example, concentrating effort on Categories 1 and 11 (90% of a 100,000-tonne footprint) and using labelled coarse estimates for small categories puts the team's effort exactly where the assurer's lands, turning closing into a clean opinion rather than a finding.