The hardest part of Scope 3 accounting for most manufacturers is not the calculation. It is getting hold of the underlying data in a usable form. Scope 3 Category 1 covers purchased goods and services, and for a mid-market manufacturer, that category is typically where the majority of total emissions sit. The good news is that you are already generating the primary data source every time you place an order: your purchase orders and supplier invoices.
This is not a complete solution. Spend-based estimates from invoices are a starting point, not a finished disclosure. But understanding exactly what emission information your purchase order stack contains, and where it runs out, is the first step toward a Scope 3 figure you can stand behind.
What a purchase order actually contains from a data perspective
A typical purchase order contains a supplier identifier, a date, one or more line items with descriptions and unit quantities, unit prices, and total spend figures. Sometimes you get a supplier part number or material classification code. Sometimes you get only a free-text description like "raw materials" or "fabrication services."
From an emissions accounting perspective, the useful fields are the line item descriptions, the physical quantities and units (kilograms, liters, units, meters), and the supplier name. The spend figures are useful only when you have no physical quantity. The total spend becomes your proxy when the line item says "machining services" and there is no other activity measure.
The description quality varies considerably across supplier relationships. A steel supplier might write "cold-rolled coil, 3mm, 2400 kg." A services supplier might write "consulting fees Q2." The first gives you an activity-based starting point. The second requires a spend-based approach by necessity, not by choice.
Why Category 1 purchased goods dominate Scope 3 for manufacturers
The GHG Protocol's Scope 3 standard covers 15 categories across upstream and downstream activities. For a typical manufacturer, Category 1 (purchased goods and services) is almost always the largest single component. The upstream production of raw materials, components, and packaging tends to carry significant embodied emissions, often more than the manufacturing operation itself.
A plastics molder buying polyethylene resin, a metal fabricator buying steel plate, an electronics assembler buying PCB substrates, each has a large chunk of Scope 3 Category 1 tied to the energy and chemical inputs of their suppliers' production processes. Those upstream emissions do not show up in the manufacturer's own energy bills; they show up in the supply chain. Which is why Category 1 is both the most important category to get right and the hardest to measure precisely.
The spend-based approach: your first pass
When you lack physical activity quantities, the GHG Protocol allows a spend-based method as an approximation. The approach multiplies total spend per procurement category by an economic input-output emission intensity factor, expressed in kilograms of CO2e per dollar of spend in that sector. These factors are derived from national economic input-output models, which are published by bodies like the US Environmental Protection Agency (USEEIO) and similar national statistics agencies.
The practical workflow looks like this: extract line items from your purchase orders, classify each by procurement category (metal, plastic, electronics, services, packaging), apply the appropriate sector-level emission intensity factor, sum across all categories. If you have twelve months of purchase orders in your ERP, you can produce a Category 1 baseline in days rather than months.
This is a genuine starting point. It is also imprecise in ways that matter for serious disclosure.
What spend-based estimation cannot tell you
The spend-based method treats all spend in a category as carrying the same emission intensity. A dollar spent on virgin aluminum and a dollar spent on recycled aluminum look identical in an EEIO model, even though their actual production emission profiles differ by a factor of ten or more.
Price fluctuations distort the calculation. If commodity prices rise, your calculated Scope 3 emissions rise even if physical volumes are flat. If a supplier shifts to a lower-cost country, the apparent emission intensity changes even if the supplier's actual process has not changed. These artifacts are real and they accumulate across a large invoice stack.
There is also the category resolution problem. EEIO models group economic sectors at a level of granularity that does not always match procurement reality. Buying industrial gases and buying contract labor sit in different EEIO sectors, but both might appear under a single spend code in your ERP. Getting the classification right at line-item level matters more than most companies expect when they start this work.
Moving toward activity-based: what your invoices give you
Activity-based calculation uses physical quantities multiplied by material-specific emission factors. A supplier invoice that says "8.5 tonnes of HDPE resin" gives you the activity data for a product-level factor rather than a sector average. The emission factor for HDPE resin production (kgCO2e per tonne) from a database like Ecoinvent or the IEA is more accurate than the general plastics sector spend factor, because it reflects the actual process rather than the average economic output of the sector.
The practical prerequisite is that your invoices actually contain physical quantities in a consistent unit. This is true for commodity inputs (steel, plastics, chemicals, fuel) more often than it is for components or fabricated parts. A steel distributor invoice almost always states tonnes. A custom machined component invoice often states only units and price, which tells you little about the underlying material mass.
Where you have physical quantities, activity-based is always preferable. Where you do not, spend-based remains your practical option, and the goal is to narrow that second group over time by improving supplier invoice specifications.
The classification problem in practice
One thing that does not get discussed enough in the Scope 3 guidance documents is the line-item classification problem. Your ERP's chart of accounts was designed by someone thinking about financial reporting, not emissions accounting. The categories that matter for one purpose are not the same as the categories that matter for the other.
"Raw materials" as a single GL code might contain steel, aluminum, plastic, adhesives, and packaging film. Applying a single average factor across that entire spend category will produce a number, but it will not be a very useful number. The work of Scope 3 data preparation is substantially the work of reclassifying invoice line items into emissions-relevant categories before you apply any factor.
Automated extraction from invoice PDFs and ERP exports can speed this classification step considerably. Parsing the line item text and matching it to a material taxonomy is something that scales. Manual reclassification of a few thousand purchase order lines per quarter does not.
A practical scenario
Consider a mid-market packaging manufacturer in Singapore buying polypropylene from three suppliers, corrugated board from two, and contract printing services from one. The polypropylene invoices state tonnes and price per tonne. The board invoices state square meters. The printing invoices state job cost only.
For polypropylene: activity-based calculation is straightforward. Tonnes multiplied by a resin production factor gives a defensible Category 1 figure that a verifier can trace to the invoice and the factor database.
For board: square meters converted to estimated mass using a standard basis weight gives an approximate activity quantity. Less precise but still an activity-based approach.
For printing services: spend-based using a print and publishing sector factor is the only practical option. The uncertainty range on this number is wider, and it should be documented as such.
Three input types, three calculation methods, all from the existing invoice stack. That is the actual texture of this work.
What this approach gives you, and what it does not
Your invoice stack gives you a defensible starting estimate for Scope 3 Category 1 that can be produced systematically and updated each quarter as new invoices arrive. For early-stage disclosure, this is a major step forward from having no Scope 3 figure at all.
What it does not give you is supplier-specific accuracy. The emission factors you are applying are averages across a production process or sector. Your specific steel supplier's actual emissions depend on their grid mix, furnace type, scrap ratio, and many other factors that a single invoice cannot tell you. Primary supplier data, obtained through direct engagement or supplier questionnaires, is the next step up in accuracy and is what most disclosure frameworks ask for as a longer-term goal.
Zevero reads invoice data to produce Category 1 estimates using activity-based factors where quantities are available and spend-based factors where they are not. We are building the data structure for Scope 3 Category 1 from documents you already have. The invoice-based figure is the starting point for supplier engagement, not a substitute for it.