Digital EPDs: machine-readable formats and why PDF is not enough

The traditional output of an EPD is a fifteen to twenty page PDF. That document is designed for a human reader and serves that purpose. But the real consumer of EPD data is increasingly software rather than a person: building LCA tools, supplier portals, public databases and procurement systems.

The structural problem with PDF

A building LCA uses hundreds of product EPDs. Reading each one's result table by hand and typing it into software is slow and error-prone. Automated extraction from PDF is unreliable because formats differ. The result is data quality degrading in transit.

The ILCD+EPD data format

ILCD+EPD is an XML-based data format developed to carry EPD data in machine-readable form. Product description, declared unit, module coverage, every indicator's value by module, data quality information and the verification record all travel as structured fields.

• Result tables transfer directly without manual copying
• Declared unit and module coverage can be checked by machine
• Validity dates are tracked automatically; expired data is flagged in the system
• Data quality and verifier information are part of the record
• Both the PDF and the structured record are generated from one dataset

Submission to databases

National databases such as ÖKOBAUDAT and program operators' own archives now accept structured datasets. The practical consequence for a manufacturer: not being able to produce the digital dataset for your EPD leaves you outside those databases, and your product is represented by a generic record in building LCA tools.

The link to the Digital Product Passport

The Digital Product Passport regulation requires a machine-readable record attached to the product, and environmental indicators are part of that record. An EPD dataset in ILCD+EPD format can feed the environmental section of that record directly. Producing digital EPDs is therefore also preparation for the next regulatory wave.