The same ton of carbon, counted more than once
A steel mill burns coal to run a blast furnace. That combustion is the mill's Scope 1 emission. The automaker that buys steel from that mill reports the same emission under Category 1, Purchased Goods and Services. The dealership that sells the finished car may report a slice of it again under its own upstream categories if it accounts for vehicle inputs. Three separate corporate inventories, one physical event.
This is not a bug in the GHG Protocol and it is not sloppy accounting. It is how the Corporate Value Chain (Scope 3) Standard is designed to work, and understanding why changes how you should use Scope 3 numbers.
Scope 1 is about control, Scope 3 is about influence
Scope 1 answers a narrow question: what emissions come directly from sources a company owns or controls. By design, no two companies should claim the same Scope 1 emission, because operational control (or in some methods, equity share) is assigned to exactly one entity. Double counting at Scope 1 would be a genuine error.
Scope 3 answers a different question: what emissions is a company connected to through its value chain, and where does it have leverage to reduce them. The fifteen Scope 3 categories — from Purchased Goods and Services and Capital Goods upstream, through Business Travel and Employee Commuting, to Use of Sold Products and End-of-Life Treatment downstream — exist to map a company's exposure and influence across a chain it does not control. A buyer cannot control a supplier's furnace, but it can switch suppliers, specify lower-carbon inputs, or use its purchasing weight to push for change. That influence is real even though the physical source of the emission sits elsewhere. Scope 3 is built to capture it.
So when a supplier's Scope 1 emission becomes a buyer's Category 1 emission, both companies are reporting something true and useful from their own vantage point. The supplier is reporting what it directly produces. The buyer is reporting what its purchasing decisions are tied to. Neither number is wrong. They are answering different questions about the same physical event.
It compounds down the chain
This layering does not stop at one tier. A raw-material producer's Scope 1 becomes a component maker's Category 1, which becomes an assembler's Category 1, which becomes a retailer's Category 1 again if the retailer accounts for the goods it sells. Downstream, a product's in-use emissions can appear in the seller's Category 11 (Use of Sold Products) and simultaneously contribute to the Scope 1 of whoever actually operates or burns the thing, if that operator also reports. The same molecule of CO2 can legitimately show up five or six times across a value chain's worth of corporate disclosures.
Why you can never sum Scope 3 across companies
This is the practical consequence that trips people up. If you add together the Category 1 figure of every company buying from a given steel mill, plus the mill's own Scope 1, plus a downstream user's Category 11, you do not get a bigger or more accurate picture of total emissions. You get an inflated number with heavy multiple-counting baked in, because the same tons are being added again and again from different standpoints in the chain.
Scope 3 totals are a per-company management and disclosure tool, not a building block for an economy-wide or sector-wide emissions ledger. National and global inventories use entirely separate accounting frameworks — generally built on direct, territorial Scope 1-style sources — precisely to avoid this layering problem. If you want to know how much total carbon a supply chain emits, you look at the primary sources (the furnaces, the fields, the tailpipes), not at the sum of every downstream company's Scope 3 report.
What this means for measurement practice
Two things follow for anyone building or reading a Scope 3 inventory.
First, treat your own Scope 3 number as internally meaningful, not comparable in aggregate to a competitor's total and definitely not addable to anyone else's. Use it to find hotspots, set reduction targets, and prioritize supplier engagement within your own boundary.
Second, data quality matters more than it might seem, precisely because Scope 3 figures travel through multiple hands. The GHG Protocol's data quality hierarchy runs from spend-based data (weakest, essentially an emissions factor applied to dollars spent) through average-data methods, to hybrid approaches, up to supplier-specific primary data (strongest, because it reflects what the upstream party actually measured). A buyer using a supplier's actual, verified Scope 1 and 2 figures to build its own Category 1 estimate is doing exactly what the system intends: passing real measured data down the chain rather than compounding generic averages. This is also where regulatory regimes like CSRD and California's SB 253, with their phased assurance requirements, are pushing: toward primary data flowing between trading partners so that each company's layered Scope 3 figure is grounded in something closer to the physical truth, even though it will still overlap with its neighbors' numbers by design.
Double counting in Scope 3 is not an accounting failure to be fixed. It reflects the fact that one ton of carbon can sit inside more than one company's sphere of responsibility at once. The framework's job is to make each of those spheres visible, not to produce a single non-overlapping global total.