Comparable or not? How dodo comparability label works

Strictly speaking, you're not supposed to compare two EPDs at all — not unless a long list of variables line up first: the same functional unit, the same Product Category Rules, the same background database, the same database version, the same characterisation factors, assumptions, and more. Meaning, almost no two real-world EPDs are ever fully comparable.

And yet companies compare them all the time. We think they're right to. 
The alternative is comparing nothing or only comparing your own progress over time. We don't believe this is conducive of a massive change we all need. 
Progress needs doing best, and knowing that needs comparison. Waiting until the whole industry is perfectly unified, every database and version and factor aligned, is waiting forever, while our rivers are drying up and heatwaves become a new norm.

So dodo takes a position: it's better to make comparisons accessible by getting the things that matter most right and letting other ones slide than to hold everything hostage to a perfection that never comes. You have to start somewhere.

Two things that keep it pragmatic rather than reckless: 
1. dodo compares only the A1-A3 stage. The production stage genuinely is in a company's control: no scenarios, minimum assumptions.

2. Scale. When two same type of products differ by a lot on the same functional unit, that gap is often real. Yes, different data base versions will produce different results, but these differences, typically, are limited. 

And perfect like-for-like isn't always the point either. A product that scores a little worse but lasts far longer (its Reference Service Life) is often a better choice. A comparison doesn't have to be flawless to point you in the right direction.

What dodo does is makes that comparison easy: you select the product, we point out whether they are comparable or not.

How it works

Where you come in

How dodo's quality label works

Every EPD on dodo platform carries a "quality" label. It's worth being precise about what it is — and what it deliberately isn't. It's a check on the data, not a verdict on the product, and not a re-assessment of the EPD: if an EPD has been third-party verified, we trust that its quality already meets a certain level. The label answers a narrower, practical question — is this EPD's data complete and internally consistent enough to compare reliably?

First, for now the algorithm checks only the machine-readable data — which can (and sometimes does) differ from what's written in the original PDF. 

Second, it checks the numbers only, because the numbers are what a comparison actually rests on, and because numbers are where errors creep in, especially mistakes introduced when data is digitised from a PDF into a database. A single mis-mapped value or a comma in a wrong place can quietly creep in and break a comparison. 

What it checks

dodo runs each digital EPD through a set of automated checks. Today there are three:

1. Are the additional impact categories present? Beyond the headline indicators, a complete record reports categories like human toxicity (cancer and non-cancer), ionising radiation, particulate matter, and ecotoxicity. Missing them isn't strictly an error — but it limits what you can compare, especially when most EPDs in the same group report the full set.
   
   
2. Does the declared climate-change total add up? The total should equal the sum of its parts — fossil, biogenic, and land-use carbon. dodo checks whether the declared total matches that sum. If it doesn't, it flags the mismatch as an error. It does not replace the number: the value you see stays exactly as the publisher declared it — dodo simply tells you it doesn't reconcile.
   
   
3. Is the declared unit valid? Some records arrive with a reference unit of zero (a stray "0 kg", "0 m²", or "1 × 0"), which makes every per-unit result meaningless. dodo flags these. 
   
   
4. Are there any significant outliers per category? This for now is a very rudimentary check that is applied on some indicators, where order of magnitude typically stays within a certain large range independently of product category or declared/functional unit (for example, ozone depletion and human toxicity, where most of the results are <0.01).

EPDs that have expired are automatically filtered out - you will not find them through the search. 

How the "alert" level is set

Each check returns a simple pass or fail, and the label reflects how many an EPD fails:

• 0 failures → OK (Green)
• 1–2 failures → Warning (Yellow)
• More than 2 → Concerning (Red)

Why it matters

A comparison is only as trustworthy as the data underneath it. Two EPDs can look comparable on the surface while one is missing half its impact categories, carries a broken unit, or has a total that doesn't reconcile. The label surfaces that at a glance — before you rely on the numbers.

What's coming

Right now the label only sees the digital data. The next step is to read the PDFs too — to catch very obvious errors. When dodo does, the goal stays deliberately narrow: check that the numbers are consistent (that the PDF and the digital data agree, and that figures add up), and check whether certain information is actually communicated. It is not a second verification. If an EPD has been verified, dodo trusts that — the check is for consistency and completeness, never a re-assessment of the underlying work.

Also in development: refined outlier detection - within a product group and the same functional unit, is a declared result per category inside the expected range, or an outlier worth a second look?

The principle behind all of it stays the same: flag what's (potentially) wrong, show why, never alter the number you're shown.

PEF Single Score - what it is and why we use it

An LCA gives you a product's environmental impact across many separate categories — climate change, water use, land use, ecotoxicity, and more. It is a complete impact, but it leaves a question: if climate change is measured in kg CO₂ and water use in m3, which one matters most for your product? That can be answered by calculating PEF Single Score.

What it is

The PEF (Product Environmental Footprint) Single Score is one number that combines all 16 environmental impact categories of the EU's Environmental Footprint method into a single figure, expressed in "points" (Pt). The point is a dimensionless unit. In short, it's a product's whole environmental profile rolled into one comparable number.

How it's calculated

The 16 categories each come in their own unit, so they can't simply be added up. Getting to one number takes three steps, using factors developed by the European Commission's Joint Research Centre (JRC):

Characterisation — each category is calculated in its own unit (climate change in kg CO₂-equivalent, ozone depletion in kg CFC-11-equivalent, and so on). This is what happens when a typical Life Cycle Assessment is calculated.

Normalisation — each result is divided by a reference value, defined by the researchers of JRC, by category. This turns every category into a unit-less number. 

Weighting — each normalised result is then multiplied by a weighting factor reflecting how important that category is judged to be. These factors, again, have been determined by the researchers. The weights sum to 100%.

Add the weighted results together, and you have the PEF Single Score.

Worth keeping in mind: the single score is a powerful summary, but it compresses all categories into one figure, so it's best read alongside the full set of impact categories, not instead of them. 

Why it's useful

First, it makes products directly comparable on overall environmental performance, not just Global Warming Potential or Carbon Footprint. which avoids the common trap of improving CO₂ while quietly worsening water or land use. 
Second, the same normalised-and-weighted breakdown that produces the score is, in fact, an instant hotspot analysis: it shows which categories actually drive a product's impact, so you know where to focus (spoiler alert: it's not always Global Warming Potential).

What data it works on

A single score can only be built on EF-compliant LCA data: results that cover all 16 EF categories, calculated with the EF characterisation factors (currently EF 3.1) and then normalised and weighted with the JRC factors. You can't derive it from a footprint that reports only GWP, or from data built on a different impact-assessment method (such as North America's TRACI) — the full EF category set has to be there.

In construction, this lines up with EN 15804+A2 EPDs, which adopted the EF characterisation factors and a broader, PEF-aligned indicator set. So A2 data is the right basis for a single score, provided all the required categories are actually reported — whereas older +A1 EPDs, built on earlier methods, generally aren't.

In dodo

dodo calculates the PEF Single Score on comparable data and lets you break it back down into the categories underneath — so you get both the one-number comparison and the hotspot view behind it. As per above, this score is only calculated on EN 15804+A2 EPD's with the full set of reported indicators. If a certain EPD has only 12 impact categories reported, you will see no Single Score displayed for that EPD. This EPD will also be generally deemed as 'not comparable' to the rest. 

An EPD is a snapshot: what it shows, and what it can't

An Environmental Product Declaration (EPD) is one of the most trustworthy pieces of environmental information a product can carry. It's verified, standardised, and detailed. But it's worth being clear about what kind of information it is: an EPD is a snapshot, taken at a point in time. 

What the snapshot captures

An EPD is built from a specific period of real production data and captures how a product was made during that period. Once verified, it's typically valid for five years.

If nothing about the product or the production process changes - the same materials, the same suppliers, the same processes, the same energy source - the EPD stays accurate for all the products made during 5 years of its validity.  There's nothing outdated about a snapshot of something that hasn't moved.

When the snapshot starts to lag

Most manufacturers, though, are changing something. And the moment production changes, the EPD's numbers begin to describe the past rather than the present.

A manufacturer is required to update an EPD before its expiry only if a reported impact gets *worse* by more than 10%. There's no matching requirement to update when things improve. So a company that has genuinely reduced its footprint can keep displaying its older, higher numbers until the EPD comes up for renewal.

Which means the direction of any error is predictable: an out-of-date EPD is more likely to overestimate the environmental impact. 

What an EPD can't show yet

Several kinds of improvement simply won't appear until the next assessment or update, even though they change the real footprint:

• Increased recycled content. Using more recycled input typically lowers a product's footprint - but a recent switch won't show in an EPD published before it happened.
• Reduced water use. Efficiency gains at the plant will only show up in the future assessment.
• A change of suppliers. Moving to lower-footprint raw material suppliers can meaningfully cut impact.
• A new take-back scheme.  Recovering products at end of life can both raise recycled content and lower the end-of-life footprint - again, not reflected until the numbers are re-run.

Where to look for the rest of the picture

If an EPD tells you where a product stood, a company's sustainability report tells you where the company is heading -  its targets, its investments, its progress on things like electrification, efficiency improvements and supplier change. The two answer different questions.

Read on their own, each is partial: a verified snapshot with no sense of direction, or a set of intentions with no verified baseline. Read together, they give you something closer to the full picture - where a product stands today, and whether the company behind it is standing still or moving.

That combination is the more accurate way to judge a product: not by a single frozen number, and not by promises alone, but by both at once.

It just might be that your product has a lower environmental impact than the one of a competitor based on a current EPD. However, what if your competitor has invested in renewable energy, is switching to 'water-less' production process or is altering a product so that it can be in use forever, with minimal repairs? That means the moment EPD is updated, the environmental impact will be nothing close to what it is now. If you don't act meanwhile and don't continue to reduce your own, you will quickly be at the bottom of the environmental performance. 

dodo merges both to give you a complete picture

That is exactly the reason why dodo is not just an "EPD comparison platform". We want to provide you with the full picture, all information necessary for you to know the state of affairs. This is why you will always see two tabs, two levels of environmental performance: a product level and a company level. 

Only by having a view on both, can a real progress be made.