Spend-based emission estimates for Scope 3 Category 1 are GHG Protocol-compliant and widely used. For most maritime operators, they also carry uncertainty of ±40–80%. Here is what that means, and what a practical path beyond it looks like.
Scope 3 Category 1: Purchased Goods and Services covers the emissions embedded in everything a vessel operator buys: provisions, stores, technical equipment, drydocking, crew services, and consumables. For a mid-sized shipowner with an annual fleet OPEX of USD 30–100 million, this is typically the largest unmeasured source of supply chain emissions.
Unlike Scope 1 (vessel fuel combustion, which is metered and relatively well understood) or Scope 2 (purchased electricity, which is modest for most operators), Scope 3 Category 1 has historically received limited scrutiny. That is changing. Three regulatory developments are pushing it up the agenda.
The EU Corporate Sustainability Reporting Directive (CSRD) requires in-scope companies to report Scope 3 emissions with increasing specificity from 2025 onwards. The European Sustainability Reporting Standards (ESRS) require companies to document their data quality and present credible improvement plans not simply report a number. The EU Carbon Border Adjustment Mechanism (CBAM), while primarily targeting direct emissions, establishes precedent for supply chain carbon accounting and signals the broader political direction. And charterer and cargo-owner requirements are becoming more concrete: large cargo owners are now embedding Scope 3 data requests into procurement and tendering processes, and spend-based estimates are no longer accepted as sufficient by leading buyers.
Most maritime operators currently calculate Category 1 using spend-based Environmentally Extended Input-Output (EEIO) methods. The approach is straightforward: multiply a supplier payment by a sector-average emission factor expressed in kg CO₂e per USD. These factors are derived from national or multi-regional input-output tables and are accepted under the GHG Protocol as a valid starting point when no better data exists.
The problem is the uncertainty embedded in this approach. A concrete example: a shipowner spends USD 1.86 million with a marine paints supplier. Applying the EEIO factor for the paints sector (0.59 kg CO₂e per USD) produces an estimate of approximately 1,099 tCO₂e. That figure says nothing about whether the paint is water-based or solvent-based, produced using renewable energy, or covered by a third-party verified Environmental Product Declaration (EPD).
Five structural factors drive the uncertainty. First, sector aggregation: a single EEIO factor covers an entire industry, meaning pumps, cranes, safety gear, and marine electronics share one emission intensity value despite having very different production profiles. Second, price inflation: input-output tables are updated infrequently, so rising prices inflate reported spend without any increase in physical volumes, causing emission estimates to drift upward independent of actual activity. Third, there is no geographic resolution: a drydock in Asia and one in Norway appear identical in a global EEIO table, despite large differences in grid carbon intensity and industrial energy efficiency. Fourth, no product differentiation: the method cannot distinguish a low-carbon product variant from a conventional one, even where a supplier has invested significantly in reducing its footprint. Fifth, currency noise: multi-currency procurement introduces exchange rate variation with no relation to physical emission volumes.
The combined effect produces uncertainty ranges of ±40–80% depending on the procurement category, based on published GHG Protocol guidance and EEIO uncertainty literature.
A recent independent analysis of a mid-sized European shipowner's Category 1 spend illustrates the scale of the issue. The dataset covered USD 46.3 million in OPEX spend across 362 active suppliers. The spend-based total came to 8,057 tCO₂e.
The breakdown by category revealed significant variation in emission intensity per dollar spent. Vessel Technical (B2) carried the highest absolute volume at 2,653 tCO₂e from USD 11.5 million spend (±55% uncertainty). Vessel Provisions (B0) produced 1,967 tCO₂e from only USD 2.4 million (±60%), reflecting the high emission intensity of food-sector EEIO factors which overweight processed goods. Dry-Dock (B8) estimated 1,966 tCO₂e from USD 9.9 million (±65%) one of the highest-impact improvement opportunities, given the large differences between Asian and European yards that the current method treats as equivalent.
Emission risk was highly concentrated: five suppliers accounted for approximately 50% of the estimated total footprint. Of the 20 largest suppliers by estimated emission volume, only 2 had publicly available product-level carbon footprint data or EPDs. The remaining 18 had corporate ESG reports, net-zero commitments, and GHG Protocol alignment statements — but no product-specific data capable of replacing the spend-based estimate.
This gap between corporate sustainability reporting and product-level PCF availability is common across the maritime supply chain. It is not primarily a reflection of intent. Product carbon footprint calculation requires methodology, tooling, and dedicated resources that most suppliers have not yet developed — often because they have not been asked with sufficient urgency, or supported with accessible tools to do it.
Replacing spend-based estimates with supplier-specific primary data produces directional shifts that are material, not marginal. For Provisions (B0), the expected shift on moving to primary data is −30 to −50%. Food-sector EEIO tables overweight processed goods, while actual maritime catering sourcing is often less processed; primary data from catering suppliers would reflect the real product mix. For Stores (B1), published EPDs from major marine coating suppliers show lower footprints than sector averages for water-based products, suggesting a shift of −20 to −40%. For Technical equipment (B2), European manufacturers typically carry lower footprints than the global EEIO averages, pointing to −25 to −45%. For Dry-Dock (B8), the gap between Asian and European grid carbon intensity is one of the most significant data quality improvements available, with an expected shift of −30 to −60%.
Applied to the case study baseline of 8,057 tCO₂e: a −30% systematic correction implies the true figure is closer to 5,600 t. A −50% correction implies around 4,000 t. The difference affects the credibility of any reduction target set against this baseline, emission intensity metrics expressed per vessel or per revenue unit, and the operator's position in charterer and investor sustainability assessments.
There is an equally important operational point: a company cannot take credit for supplier decarbonisation it cannot measure. If a key paint supplier switches to a lower-carbon formulation, or a drydock installs renewable energy capacity, the spend-based method will never register it. The numbers simply do not move.
Moving to primary data does not require engaging every supplier simultaneously. Concentration analysis makes the task tractable. In the case study described above, engaging fifteen to twenty suppliers would cover 70–80% of estimated Category 1 emissions with primary data, while the remaining tail continues on spend-based methods during the transition.
Effective supplier engagement follows a clear sequence. Prioritise by emission weight, not spend — the largest invoice is not always the largest emission contributor. Request data in a standardised format, such as PACT Pathfinder 2.0 or ISO 8266-compliant PCF declarations, to enable aggregation rather than bespoke back-and-forth exchanges. Provide accessible tooling rather than simply issuing a data request: suppliers engage more readily when the process is clear. Start with suppliers already active on sustainability, as they are most likely to have partial data and to respond constructively. Sequence engagement with critical suppliers first, high-impact relationships second, and broader rollout over a twelve to eighteen month horizon.
The GHG Protocol acknowledges that Scope 3 data quality improves over time and that the key obligation is transparency about current quality combined with a documented improvement plan. A transition from spend-based to supplier-specific data — with clear milestones and expanding primary coverage — is both defensible to auditors and demonstrably better than a static spend-based position. Proposals to the GHG Protocol Scope 3 revision process (2024 consultation) include calls to phase out spend-based methods as a long-term primary approach. Operators who begin the transition now are building a compliance advantage, not just improving accounting accuracy.
If your current Scope 3 Category 1 figure rests entirely on spend-based EEIO estimates, four questions are worth addressing before the next reporting cycle. How concentrated is your emission footprint — if five or ten suppliers drive the majority of your estimated emissions, the improvement path is tractable. How material is the uncertainty in each category for your reporting obligations. Which of your largest suppliers are most likely to have, or be able to produce, primary PCF data. And what credible improvement plan can you document for your CSRD disclosure or for charterers and investors who are beginning to ask.
ReFlow offers a free Scope 3.1 hotspot analysis: a structured assessment of where your Category 1 data carries the most uncertainty and which supplier relationships to prioritise first. To request one, contact us.