Where a First EPD Cycle Is Won or Lost: A Verifier's Notes from Jordan

A project started in Amman on 29 September: the Jordan Green Building Council, with support from GIZ’s GAIN programme, is placing ten trainees who have completed LCA training inside five industrial plants, targeting five LCA studies and two internationally published EPDs. We summarised the launch itself on our news page; this article looks at the project from a verifier’s chair, as the person who gave the opening’s technical session: what has to be set up right in a first cycle, where time gets lost, and what “environmental data for exports” means in practice.

Why “capacity” is the right word

What stands between a manufacturer and an EPD is rarely software or arithmetic. It is local people who can run the study and data that is ready for verification. An external consultant produces one EPD and leaves; if no data-collection discipline stays behind in the plant, the second EPD costs more than the first. The project’s design — trainees embedded in plants — aims at exactly this. Ten people collecting data on five real production lines learn what no seminar can teach: where the data sits, who holds it, and which number is unreliable and why.

Jordan’s Minister of Environment framed it the same way at the launch: documented environmental data is no longer optional for factories selling into Europe. The word to underline is “documented”. A figure on a slide and a declaration that has passed independent verification are not the same thing.

The first ninety days for five plants

In a first LCA cycle the decisive choices come before the calculation. In our experience these have to be closed in the first ninety days:

• PCR and standard version. For construction products in the European market, the core standard is EN 15804+A2, with complementary PCRs (c-PCRs) for some product groups. Collecting data before the PCR is chosen is the most expensive mistake: the declared unit, module scope and scenario rules come from the PCR, and if it changes later, part of the collected data is wasted.
• Declared unit and system boundary. One tonne, one square metre, one cubic metre? Which modules are required? Module scope under EN 15804+A2 is wider than under +A1; a study planned on an “A1-A3 only” assumption comes back from verification.
• Primary data list. The plant data needed for A1-A3 (raw material quantities, energy, water, waste, transport distances, production volume) and the reference year should be defined at the start, and the source of every figure (meter, invoice, stock record) recorded.
• Allocation rule. In multi-product plants — and at least some of the five pilots are — the allocation approach should be written down from the PCR at the outset, not improvised at the end.
• Background data and characterisation. The database, its version and the EF-based characterisation set required by +A2 should be fixed at the start. Putting the figures next to an EPD of the same product prepared under +A1 is meaningless; we covered this in the guide to reading an EPD.

The list looks long, but none of it is a new requirement; these are the questions verification already asks. Answering them at the start prevents re-collecting data at the end.

Publishing two EPDs internationally: what it takes and what it does not

One of the targets is for two plants to publish their EPDs internationally. The process is clear and cannot be shortened: the EPD is prepared to a programme operator’s PCR, passes independent third-party verification under ISO 14025, and is registered in the operator’s system. Which operator to choose depends on the target market, the product group and the operator’s PCRs. There is no “best operator”; there is the operator that fits this product and this market.

It also helps to describe verification accurately. The verifier confirms that the declaration conforms to the PCR, EN 15804 and ISO 14025 and that the data is traceable; the verifier does not decide that the product is “green”. An EPD is a standardised information card, not a performance grade. That is why a target of two EPDs in a first cycle is a measured one: verification expects every figure from the plant to have an answer to “where did this come from?”, and that answer usually takes a few rounds the first time.

EPD data and CBAM data are not the same thing

The topic that drew the most interest in the opening session was the relationship between the Carbon Border Adjustment Mechanism (CBAM) and EPDs. Some of the pilot plants — aluminium and fertilisers — work in product groups within CBAM scope. Two misunderstandings are worth clearing up early:

1. An EPD does not replace a CBAM report. The EPD reports life-cycle indicators (including GWP) per declared unit under EN 15804. CBAM asks for the product’s specific embedded emissions under its own implementing rules, with its own system boundary and unit. The boundary, allocation rules and scope differ; a GWP-A1-A3 figure from an EPD cannot be copied into a CBAM declaration.
2. But the same plant data feeds both. Production volume, fuel and electricity consumption, input materials and their source records — collected properly once, they become inputs to two separate calculations. This is the project’s real gain: the data discipline the trainees build carries not only the plants’ EPDs but their export reporting as well.

So we read “environmental data is mandatory for exports” this way: what is mandatory is not a single document but traceable plant-level data. The document is the output of the data.

Five notes from a verifier’s chair

Short notes for the project teams — and for any manufacturer starting a similar first cycle:

1. Build the data trail first. Every figure should carry its source, date and owner. The most frequent question in verification is “where did this number come from?”
2. Leave missing data missing; do not fill it in. If an unknown transport distance or density must be estimated, state that it is an estimate and what it rests on. A silently assumed value is the finding that costs the most time in verification.
3. Do not leave modules C and D for last. End-of-life scenarios and recycling assumptions should be planned early according to the PCR; “production is done, now let’s write the scenarios” produces inconsistencies.
4. Do not compare — yet. A first EPD is made to build the plant’s own data base, not to race other EPDs in the sector. Comparison is meaningful only with the same PCR, the same declared unit and the same module scope.
5. Put verification time in the calendar. Verification is a review round, not a signature; questions come back and data gets reopened. For an April 2027 target, at least a few months should be reserved for it.

Closing

The project aims at the right place: people and data. Five plants and ten trainees are building what Jordan has long lacked — local LCA/EPD practice. The sources behind the opening talk, Jordan’s export and CBAM context and a glossary of the terms used are on our kick-off page. For manufacturers preparing a similar first cycle, what an EPD is and how the process runs is summarised in a separate article.