Spend-Based Emissions: The Starting Point Most Australian Companies Actually Need

We ran the same procurement dataset through three different spend-based emission factor databases last month. The results: 14,200 tonnes, 9,800 tonnes, and 22,100 tonnes. Same invoices. Same dollar amounts. Three wildly different answers.

That's the reality of spend-based emissions accounting. And it's still the method most Australian companies will use for their first Scope 3 disclosure.

Not because it's good. Because it's possible. The GHG Protocol is clear: supplier-specific data beats hybrid data beats spend-based data. But when your Scope 3 reporting deadline arrives and you have procurement spend data in your ERP but nothing resembling activity data from 200 suppliers, spend-based is what you've got.

The honest take from our side: spend-based emissions are a floor, not a ceiling. They're how you figure out which categories matter. They get a number into your ASRS disclosure that's defensible (if documented properly) under the modified liability protections running through December 2027. But they're not where you stay.

What Spend-Based Actually Means

The formula is almost laughably simple. You take the dollar amount you spent with a supplier or on a procurement category, and multiply it by an emission factor expressed in kilograms of CO2-equivalent per dollar.

Spend ($AUD) x Emission Factor (kgCO2e/$AUD) = Estimated Emissions (kgCO2e)

That's it. No physical activity data. No kWh, no litres, no tonnes of material. Just money in, emissions out.

The emission factors themselves come from environmentally extended input-output (EEIO) models. These models take national economic data — how much money flows between industries — and overlay environmental accounts showing how much each industry emits. The result is an average emissions intensity per dollar of output for each industry sector.

The main global databases are EXIOBASE (8,000+ entries across 49 regions, denominated in EUR), USEEIO (US EPA, denominated in USD), and UK DEFRA/DESNZ factors (denominated in GBP). For Australian companies, there's also the Australian Spend-Based Emission Factor dataset published through Zenodo, derived from ABS input-output tables coupled with EXIOBASE data, expressed in kilotonnes CO2-e per million AUD.

Here's the wrinkle that catches most people. DCCEEW doesn't publish spend-based factors. The NGA Factors workbook that Australian companies rely on for Scope 1 and 2 covers physical activity — kWh, GJ, litres, tonnes. It has nothing to say about dollars. So you're immediately working with international datasets, currency conversions, and emission factors calibrated to economic structures that don't quite match Australia's.

The Currency Problem Nobody Talks About

Most EXIOBASE factors are in EUR. USEEIO is in USD. DEFRA is in GBP. Your invoices are in AUD. Some of your invoices — international freight, imported materials, overseas IT services — might be in USD or other currencies.

This creates a layered conversion problem. You need to convert your spend into the currency that matches your emission factor, apply the factor, and get emissions out the other side. Get the exchange rate wrong and your numbers move by 10-20% before you've even thought about methodology.

And it gets worse. Exchange rates fluctuate. If you're applying 2022 EXIOBASE emission factors to 2026 procurement data, which exchange rate do you use? The rate from 2022 (matching the factor vintage)? The rate from 2026 (matching the spend data)? The annual average? There's no universally agreed answer, though using the factor vintage year makes the most methodological sense.

In Carbonly, we handle this with built-in currency conversion across AUD, USD, EUR, GBP, NZD, CAD, and SGD. When an emission factor is denominated in USD but your invoice is in AUD, the system converts automatically using the appropriate rate. Our AI document extraction recognises currency patterns in any format — $45,000, A$45,000, AUD 45,000, US$12,000 — and maps symbols to the correct currency. It sounds like a minor detail until you're processing 500 invoices across three currencies and realising that "

quot; on an Australian invoice means AUD but "

quot; on a US supplier's invoice means USD.

Where Spend-Based Accuracy Falls Apart

We need to be blunt about this. Spend-based is the least accurate method in the GHG Protocol hierarchy, and the error margins are large.

NQA analysed over 50 ISO 14064-1 verified organisations across 28 industry sectors and found that 77% of them had actual emissions lower than what spend-based factors predicted — on average, 79% lower. The 23% that exceeded the factors? They were higher by an average of 623%. That's not a rounding error. That's an order-of-magnitude problem in some sectors.

Research cited by Climatiq shows EEIO factors can carry uncertainty ranges of plus or minus 50-200%. Compare that to product-specific lifecycle factors, which typically sit at plus or minus 25-50%. A study by Steubing et al. found that over half of product footprints differed by a factor of two when comparing EEIO-derived results against process-based LCA data.

Three specific failure modes hit Australian companies:

Inflation distortion. Spend-based factors are calibrated to a specific price year. Most available factors use 2019-2022 economic data. If you're applying them to 2026 spending without deflating for inflation, your emissions estimate goes up every year even if your actual procurement volumes are flat. Australia's CPI rose roughly 20% between 2019 and 2025. That's a 20% emissions overstatement before you've made a single methodological choice. The proper fix is deflating spend to the factor's base year using CPI or producer price indices. Most companies don't do this.

Supplier indifference. A spend-based factor treats every dollar spent in an industry sector the same. Your concrete supplier running on 100% renewable electricity looks identical to the one burning coal. A supplier that's cut emissions 40% through efficiency improvements gets the same factor as the worst performer in their sector. The method is structurally blind to supplier effort — which, ironically, removes the incentive for suppliers to improve.

Sector classification mismatch. EXIOBASE uses its own industry classification. USEEIO uses US NAICS codes. Australian procurement systems use UNSPSC or internal categories. Mapping your spend into the right emission factor sector is subjective, and different mappings produce different results. That's a big part of why our test with three databases produced three different answers.

When Spend-Based Is the Right Call

Despite everything above, spend-based emissions accounting has a legitimate and important role. The GHG Protocol explicitly permits it. AASB S2 accepts estimation methods where "reasonable and supportable information [is] available without undue cost or effort." And the modified liability framework protects Scope 3 disclosures (including methodology choices) through December 2027, with only ASIC able to take enforcement action.

Spend-based works well for specific situations.

Initial screening. Before you invest months collecting supplier data, run a spend-based estimate across all fifteen Scope 3 categories. It'll tell you that purchased goods and services is 70% of your footprint and employee commuting is 2%. That stops you wasting effort on categories that don't matter.

The long tail. Even companies with mature Scope 3 programs use spend-based for their smaller suppliers. Your top 20 suppliers by spend might provide activity data. Your next 200 probably won't. And they probably don't need to — the emissions from supplier number 150 aren't going to change your materiality assessment.

Categories where activity data doesn't exist. Some Scope 3 categories are inherently hard to measure with physical data. Category 2 (capital goods), portions of Category 1 (purchased services like legal, accounting, IT), and some aspects of Category 4 (upstream transport arranged by suppliers) often lack activity-level data. Spend-based may be the only practical option.

Year-one reporting under ASRS. Group 2 entities starting from July 2026 get a one-year Scope 3 deferral. When they hit their second year, a well-documented spend-based estimate is far better than an empty disclosure. The auditor won't love it. But under ASSA 5010's limited assurance requirements for early-year reporting, a documented methodology with traceable source data and acknowledged limitations is acceptable.

In Carbonly, our material matching system handles this automatically. When the AI extracts a monetary amount from an invoice — say, $45,000 for "IT consulting services" — the Tier 5 LLM matching engine identifies this as a spend-based item, finds the appropriate currency-based emission factor (kgCO2e/AUD), and calculates emissions directly. It also generates context keywords — "professional services," "IT services," "consulting" — to ensure the factor matches the actual service, not just the broad sector. That matters because the difference between "IT services" and "data centre hosting" can be a factor of five in emission intensity.

The Path from Spend-Based to Activity-Based

Here's our actual position: every Australian company doing Scope 3 should start with spend-based estimates and plan to move away from them.

The GHG Protocol describes a data quality hierarchy. From worst to best:

Spend-based — dollar amounts multiplied by EEIO factors
Average-data — physical quantities multiplied by industry-average emission factors
Hybrid — mix of supplier-specific data and secondary data
Supplier-specific — actual emissions data from your suppliers, allocated to your purchases

You don't have to jump straight to supplier-specific. The realistic path for most Australian companies over the next two to three years looks like this:

Year one: Spend-based screening across all categories. Identify your top three to four material categories. Document your methodology in a Basis of Preparation memo. Start conversations with your top 20 suppliers about providing activity data.

Year two: Shift your top five to ten suppliers (by emissions contribution) from spend-based to activity-based, using their electricity bills, fuel records, and transport data. Keep spend-based for the long tail. Your auditor under ASSA 5010 will want to see documented improvement in data quality year-on-year.

Year three and beyond: Expand activity-based coverage. Pursue supplier-specific emission allocations from your largest partners. The hybrid method — real data where you have it, estimates where you don't — is where most companies will realistically land.

We're not sure this approach fully works for companies with 500+ suppliers across multiple countries yet. The data collection logistics are genuinely unsolved at that scale for mid-market companies without dedicated procurement sustainability teams. We're building tools to help, but we won't pretend the problem is easy.

What Your Auditor Will Actually Ask For

Even though Scope 3 disclosures are protected by modified liability through December 2027, your assurance provider will still review them under ASSA 5010's limited assurance requirements. Here's what the engagement partner will want to see, based on what we've observed from Group 1 entities' first reports:

A Basis of Preparation document explaining your methodology for each material Scope 3 category. This should state which emission factor database you used, why you chose it, how you handled currency conversion, whether you adjusted for inflation, and what assumptions you made about sector classification.

Source data traceability. Can you trace a Scope 3 number back to a procurement export, a supplier invoice, or a general ledger entry? If your spend-based calculation is a single spreadsheet formula pointing at a GL total, that's thin. If it's individual line items mapped to emission factors with source documents attached, that's audit-ready.

A data improvement plan. Auditors expect to see a roadmap showing how you'll move from spend-based toward activity-based or hybrid methods over time. AASB S2's "reasonable and supportable information" principle isn't a permanent hall pass — it's a recognition that data quality improves year on year.

Consistent category boundaries. If you included business travel in this year's Scope 3 but excluded it last year, you need to explain why. Similarly, if you changed emission factor databases between years, the auditor will want to understand the impact on comparability.

In Carbonly, every spend-based calculation carries a full audit trail — the source invoice, the extracted amount, the currency detected, the conversion rate applied, the emission factor used (with database and vintage year), and the final calculation. That's the kind of documentation that makes an assurance engagement survivable instead of a scramble.

A Worked Example: Mid-Market Construction Company

Consider a mid-market construction company with $85 million in annual procurement spend across roughly 180 suppliers. Their Scope 1 is diesel and gas. Their Scope 2 is electricity. But their Scope 3 Category 1 — purchased goods and services — is where the real number lives.

Using the Australian spend-based emission factors from the Zenodo dataset, here's a rough breakdown of what their procurement looks like:

Procurement Category	Annual Spend	Approx. Factor (tCO2e/M AUD)	Estimated Emissions
Concrete and cement products	$22M	~600	~13,200t
Steel and metal products	$15M	~800	~12,000t
Electrical and plumbing services	$12M	~200	~2,400t
Plant hire and equipment	$10M	~350	~3,500t
Professional services (engineering, surveying)	$8M	~80	~640t
IT, admin, office supplies	$5M	~100	~500t
Other (insurance, finance, telecoms)	$13M	~60	~780t

Total estimated Scope 3 Category 1: roughly 33,000 tonnes.

That's a useful number. Not because it's precise — the actual figure could be anywhere from 20,000 to 50,000 tonnes depending on factor choice and classification. But because it tells you where to focus. Concrete and steel are 76% of this estimate. That's where you invest in getting activity data — actual tonnes purchased, actual embodied carbon from EPDs, actual supplier emission reports. Professional services and IT? Leave them on spend-based. Moving those from spend-based to activity-based won't materially change your total.

What Spend-Based Can't Do

A few things that catch people out.

Spend-based can't show year-on-year emissions reduction unless you're physically buying less stuff. If your concrete supplier switches to a lower-carbon mix but charges the same price, your spend-based number doesn't change. If they charge more for the low-carbon product, your spend-based estimate actually goes up — the exact opposite of what happened physically. This creates a perverse signal where the greener option looks worse.

Spend-based can't support product-level carbon footprinting. If a customer asks for the carbon intensity of a specific product you sell, you can't answer that question with spend-based data. You need lifecycle assessment, and that requires physical data.

Spend-based won't survive tightening assurance forever. ASSA 5010 moves from limited to reasonable assurance for financial years starting from 1 July 2030. Reasonable assurance means auditing 80-90% of your evidence, not reviewing 10-20%. An entire Scope 3 inventory built on spend-based estimates will get uncomfortable scrutiny under reasonable assurance. The expectation — even if not yet written into the standard — is that methodology quality improves over time.

And spend-based doesn't satisfy the spirit of what AASB S2 is trying to achieve. The standard links Scope 3 to transition planning, climate risk assessment, and capital allocation decisions. If your Scope 3 number changes by 30% depending on which emission factor database you pick, it's hard to build a credible transition plan on top of it.

Getting Started Without Overthinking It

If you're a Group 2 entity staring down your first reporting year and haven't touched Scope 3 yet, here's what we'd do.

Pull your procurement data from your ERP or accounting system. Sort by spend. Classify your top 50 suppliers by industry sector — use ANZSIC codes if you have them, or just describe what they do. Map those sectors to the closest EXIOBASE or Australian EEIO category. Multiply. Write down every assumption you made while doing it.

That's your screening estimate. It'll take a few days, not months. And it'll tell you which three categories to focus your real data collection effort on.

Then start the harder work. Ask your top ten suppliers for electricity bills and fuel records, not emissions reports. Most of them can produce a twelve-month electricity summary within a week. That activity data, combined with NGA Factors, gives you a number that's ten times more accurate than the spend-based estimate — and it's genuinely not that hard to get.

Spend-based is a starting line. A useful one. But not a place to stay.

Related Reading: