Averaging and Grouping Rules for HPDs vs. EPDs

Why grouping even matters

Grouping trims admin work, speeds specs, and keeps the story consistent across marketing and submittals. It also protects credibility, because reviewers now spot sloppy averaging in minutes. Think of it like a playlist that actually matches the vibe of the room, not a random shuffle.

EPDs: the 10% leash on impacts

Most construction PCRs under the International EPD System allow a single EPD to cover multiple similar products only when the declared impact indicators stay within about 10% across included variants if the EPD claims ISO 21930 compliance (EPD International FAQ, 2025). If the spread is larger, you either justify the grouping and declare the variation, or switch to worst case or separate EPDs, and you must present only one set of results per EPD under GPI 4.0 rules (EPD International, 2025). In practice, check A1 to A3 first, then confirm other indicators so the whole set does not blow past that 10% threshold (EPD International FAQ, 2025).

HPDs: similarity of recipe first, numbers second

HPDs focus on what is inside the product. One HPD can cover multiple SKUs when content is identical or when differences are up to roughly 10% of total mass, and v2.3 explicitly allows larger differences with a clear explanation of how and why the formulations diverge (HPDC v2.3 FAQ, 2025). The frame is disclosure precision, not life cycle impacts, so screens and thresholds like 1000 ppm or 100 ppm matter more than kilograms of CO2e (HPDC Manufacturer Guide, 2018).

The reference configuration for modular systems

Custom sizes, cut lengths, stack‑ups, or thickness options need a reference configuration that real buyers actually choose. For EPDs, pick a representative or volume‑weighted product and show how far the rest of the family sits from that anchor, then report the variation as required if any indicator drifts toward the line (EPD International FAQ, 2025). For HPDs, choose the base bill of materials, document alternates, and explain any percent‑by‑mass differences beyond the typical 10% window in plain language (HPDC v2.3 FAQ, 2025).

Three knobs when EPD variance is high

If modeling shows indicator spreads above 10%, you still have options. Worst‑case EPDs are allowed and remove averaging risk, at the cost of conservatism that may sting on competitive projects (EPD International FAQ, 2025). Parameterized EPDs can work where the PCR and program allow them, but they need disciplined inputs, consistent cut‑offs, and clean guardrails so the math stays defensible.

Keep HPD and EPD data in lockstep

Start with an identical BOM spine for both documents. Align suppliers, lot‑specific alternates, and adhesive systems, then translate that spine into LCA data on one side and hazard screening on the other. Do not let the HPD list ingredients the EPD never modeled in A1 to A3, or the other way around, since reviewers will notice and ask questions fast.

Decision shortcuts that save weeks

• If SKUs share the same recipe and only size changes, one HPD and one EPD usually suffice. Confirm A1 to A3 stays tight on mass‑normalized results for the EPD, and document identical content for the HPD.
• If the recipe changes in small ways, one HPD can still work with clear notes on alternates and mass differences. Check the EPD spread across all indicators before averaging.
• If the recipe changes materially, split the HPD or separate the content into clearly related families. For the EPD, choose worst case, parameterize with guardrails, or publish separate EPDs.

Common pitfalls to skip

Publishing an averaged EPD without disclosing the spread invites corrections. Treating a colorant package as a minor tweak when it shifts binder content is another classic miss. And assuming a buyer will do conversion math from a base EPD to a custom configuration is not allowed by program rules, so do not go there (EPD International FAQ, 2025).

What to document so reviewers say yes

State whether your EPD represents average, representative, or worst case, and disclose any distance between the declared value and the farthest SKU if required by the PCR (EPD International FAQ, 2025). In HPDs, call out alternates, note residuals work, and keep thresholds explicit so LEED reviewers can check the box without emails back and forth. It sounds straghtforward because it is, when the BOM is stable.

Wrap up without the hand‑waving

A single declaration per family is efficient when the numbers and recipes truly cluster. EPDs hold the line on impact variation for comparability, HPDs reward transparent formulation logic. Nail the reference configuration, decide early when to separate or to go worst case, and keep A1 to A3 inputs synchronized with your HPD content. Teams that do this spend less time arguing spreadsheets and more time winning specs.