Dataset selection is one of the most consequential modelling decisions in a life-cycle assessment.
A calculation may be technically correct yet still produce misleading results if the chosen data represents a different product, production route, geography, time period or methodology. A clear selection process strengthens credibility, consistency, verification readiness and auditability.
This guidance presents a practical four-tier workflow for each material, process or component modelled in One Click LCA. It is informed by the data-quality and representativeness principles referenced in EN 15941:2024 and EN 15804+A2. It supports professional judgement; it does not replace mandatory programme-operator, national, certification, PCR or calculation-tool rules.
Core rule: Use the most specific source that is both eligible and representative. Move to the next tier only when the higher tier is unavailable or unsuitable. |
Figure 1. Decision flow for selecting the first eligible and representative source.
1. Follow the four-tier hierarchy
Begin with the strongest source that accurately represents the item being modelled. Verification status alone is not enough: a higher-tier source should be rejected when its product scope, manufacturing technology, geography, validity or methodology does not fit the assessment.
Tier 1 — Verified product or study data
1a. Manufacturer-specific Type III EPD- use a valid, independently verified EPD (ISO 14025) for the actual product and relevant manufacturing route (supplier specific) whenever it is available and eligible.
1b. Manufacturer-average or product-family EPD - use when the declared product group, plants and averaging approach reasonably cover the modelled item.
1c. Collective or sector EPD- use when it represents the relevant market, product class and production technology.
1d. Critically reviewed LCA study - use when no suitable EPD exists and the study is particular to the product and follows an appropriate LCA standard, system boundary and critical review process.
Important: A verified source from the wrong product, market or technology can be less representative than a suitable generic dataset. |
Tier 2 — Generic LCI dataset
Use a generic life-cycle inventory dataset when no Tier 1 source adequately represents the material or process. Select it by geography, production technology and time. Availability and permitted use may depend on the selected database licensing and costing.
Tier 3 — One Click LCA generic profile
Use a One Click LCA generic profile when verified sources and suitable background LCI datasets do not cover the material or process. Review the dataset description, geographic scope, modelling assumptions and limitations before use.
Tier 4 — Documented proxy
Use a proxy only as a last resort. Select the closest defensible material or process by function, composition, performance, technology and geography. Record the substitution, any adjustments and the resulting uncertainty. Never select a proxy because it gives a lower impact result.
Figure 2. Four-tier hierarchy for environmental data selection.
2. Confirm representativeness before accepting the dataset
Every shortlisted source regardless of tier should pass all three checks below.
Geographical — Does the dataset reflect the manufacturing location or supply chain? Prefer the narrowest valid geography; use regional or global data only when a more specific source is unavailable or misleading.
Technological — Does it match the production route, feedstock or recycled content, energy mix, performance class and co-product treatment? If the route is unknown, use a market average rather than a best case.
Time and validity — Is the EPD valid and the reference period sufficiently current? Use the latest supported database release and document any justified exception.
Figure 3. Four checks for geographical, technological, temporal and methodological fit.
3. Choose among generic LCI databases
There is no universal ranking of generic LCI databases. First apply mandatory national, programme-operator, PCR specific requirements. Among the eligible options, choose the database whose sector, geography and technology best fit the project.
Category | Use when | Examples (non-exhaustive) |
General-purpose | Broad material, process and geographic coverage is needed. | BAFU, Ecoinvent, IDEMAT, ProBas |
Sector-specific | Specialised process or supply-chain detail materially improves representativeness. | Carbon Minds, Worldsteel, PlasticsEurope, WEEE, CEPE |
Programme / national | A national method, PCR or programme operator prescribes a source. | KBOB / BAFU, NMD / MRPI and other scheme-prescribed datasets |
Database names are examples, not a ranking. Confirm current availability, licence access and permitted use in the selected One Click LCA calculation tool.
When two eligible datasets are equally representative: prefer the project’s primary background database to preserve consistency; use the latest supported version; and document any reason for an exception.
Figure 4. Selecting a generic LCI database and resolving equally suitable options.
4. Record the selection
For every bill-of-materials item, record enough information for another practitioner or verifier to understand and reproduce the decision.
1 | Dataset identity. Name, provider or database, and a stable reference where available. |
2 | Selection rationale. Tier used, why the source represents the item, and why a higher tier was unavailable or unsuitable. |
3 | Version and time. Database version, EPD publication or validity date, reference year and relevant geography. |
4 | Assumptions and limitations. Adjustments, proxy substitutions, exclusions and material uncertainties. |
5 | Applicable rule. Calculation tool, programme, national method, PCR or certification requirement that affected the choice. |
Outcome: Consistent documentation makes the model auditable, supports verification and reduces the risk of untraceable changes during project updates. |
Standards and related guidance
Standards referenced: EN 15941:2024; EN 15804:2012+A2:2019; EN ISO 14025; EN ISO 14044; ISO 21930; and the applicable PCR, programme-operator and national rules.
Before publication, confirm clause references against the applicable licensed standard and current programme or national requirements.