EcoVadis rejected documents: why good evidence still loses points

EcoVadis rejects supporting documents for a few predictable reasons: the document looks informal or undated, it was written only for the questionnaire, it bundles unrelated evidence into one file, it is out of date, or it does not cover enough of the business being assessed. Most lost points come from how evidence is packaged. The underlying work is often fine.

A strong company, a falling score

We spoke to a large services business recently that had held an EcoVadis gold medal for two years running, then silver, then watched its percentile drop from the mid-90s to the low 60s in a single cycle. Nothing about how it ran health and safety, training or procurement had got worse. What changed was the scope: two European entities were added to the assessment, and the evidence that had worked for the UK business no longer covered the whole group.

The scorecard came back with 35 high and medium improvement areas. One training document had been rejected for five separate reasons at once. Others were marked as "created for EcoVadis", even though they had been written to act on the previous year's feedback. The team spent hours copying rejection comments from nested screens into a spreadsheet, because there was no single export that showed every document and why it failed.

That last detail is the one I keep thinking about. A company that scores at the top of its own industry's regulatory audits could not even see, in one place, why it had lost points. If that is where your team is today, the problem is very fixable, but you need to start with the rules EcoVadis actually applies.

Why EcoVadis rejects documents

EcoVadis publishes its own guidance on supporting documents, and most rejections map straight onto it. These are the reasons we see most often on scorecards:

• Informal documents: EcoVadis does not credit documents without the company's name or logo, or without an issue or review date. Photos and screenshots usually fall into this bucket too.
• Created solely for the questionnaire: documents produced only to satisfy the EcoVadis questionnaire are not accepted, and the analyst comment appears on the scorecard, where requesting customers can see it.
• Combined documents: EcoVadis says it does not accept a collection of unrelated evidence, such as charts, receipts and photos, merged into one file.
• Out of date: EcoVadis applies a validity period of 8 years to policies and actions, and 2 years to KPI reporting. Older data is discounted.
• Wrong scope or too little coverage: evidence has to relate to the scope being assessed. In its Q1 2026 methodology update, EcoVadis said site-level evidence must cover at least 80% of operations, and 95% for energy and greenhouse gas KPIs.
• Links instead of files: analysts will not download documents from embedded web links, so a policy that only lives on your intranet earns nothing.
• Not linked to a criterion: evidence that does not relate to one of the 21 EcoVadis sustainability criteria is not credited.

The "created for EcoVadis" trap

This is the rejection that frustrates people most, and I understand why. EcoVadis tells you what is missing, you fix it, and then the fix is rejected for being made for EcoVadis.

The way through is to make sure every new document lives inside your normal management system before it goes anywhere near the platform. A training record should come out of the system you use for all training, carry the same branding and version control as every other record, and show a reporting period of at least a year. A policy should be approved, dated and published to staff, ideally a few weeks before you submit. If a document could only exist because of the questionnaire, an analyst will probably spot it.

EcoVadis does allow some documents that look combined. Its guidance accepts certificates for multiple sites, consolidated KPI reporting for multiple sites, a training plan bundled with its materials and attendance records, and policy subsections inside a management manual. Merging four unrelated PDFs into one upload to save attachment slots is a different thing, and it gets marked down.

Why adding entities can drop your score

Group restructures are where scores fall hardest, and the coverage rules explain most of it. When new entities join the assessment, every piece of site-level evidence has to stretch across them. A UK-only training matrix, ISO certificate or energy dataset suddenly covers a smaller share of operations, and if it falls below the coverage thresholds the evidence is rejected rather than partly credited.

The fix is unglamorous: build a simple matrix of every criterion against every entity before you submit. For each cell, note the document, its date, its reporting period and the share of employees or sites it covers. Gaps show up immediately, and you can decide whether to extend a policy group-wide, collect local KPIs or accept a lower score on that criterion this cycle.

Medals make this worse because they are relative. EcoVadis awards platinum to the top 1%, gold to the top 5%, silver to the top 15% and bronze to the top 35%, compared with every company assessed in the previous 12 months. You can stand still and still lose a medal as the cohort improves.

How to fix rejected evidence before your next assessment

1. Export every rejection reason into one list. Include low-priority items as well as high and medium. Group them by reason, not by question, because the same root cause usually repeats.
2. Map evidence against entities and criteria. Check coverage for each criterion, especially energy and greenhouse gas KPIs, where the threshold is 95%.
3. Replace informal documents. Add company branding, issue dates, owners and review dates through your normal document control.
4. Split combined files. One document per piece of evidence, unless it fits one of the accepted exceptions.
5. Refresh KPI data. Anything older than 2 years needs a newer reporting period with clear units and a stated year.
6. Plan your 55 documents. EcoVadis sets a limit of 55 documents per assessment regardless of company size, and 30 MB per file. Spend them on the criteria with the most points at stake.
7. Use the factual error review. From Q2 2026 EcoVadis offers a two-business-day window to flag factual errors before a scorecard is published. Have someone ready to check it.

Our guide to improving your EcoVadis score covers the wider plan, and what is a good EcoVadis score explains how the percentile bands work in practice.

The data behind the documents

Many rejections are really data problems wearing a document costume. "Reporting unit missing", "data over two years old" and "period covered less than a year" all point to KPIs that were pulled together by hand from colleagues in different countries. Environment questions are the clearest example. If your energy and emissions figures do not cover the whole group, with units, periods and sources, the evidence falls short however well the policy is written. That is why a proper Scope 1, 2 and 3 inventory matters for EcoVadis as much as it does for a net zero target, a point we cover in carbon neutral vs net zero.

It is also why one person owning everything tends to break at exactly this moment, something we wrote about in ESG data management when one person owns it.

Where Ella fits

Ella is AI-native ESG software built as an agent, with a dedicated EcoVadis module. It reads your scorecard and shows where you are losing points, then does the fixes: it drafts missing policies and evidence in your own format and collects the KPI data from colleagues in each entity, chasing data owners and tracking missing responses until every number has a source. Expert support from Ella's team sits behind it for the judgement calls. If your last scorecard left you with a long list of rejected documents, see how Ella improves EcoVadis scores.