How monitoring bodies test accessibility: method, samples and results

Which legal basis requires monitoring?

The monitoring obligation is laid down in Art. 8 of Directive (EU) 2016/2102: Member States periodically monitor whether websites and mobile applications of public sector bodies meet the accessibility requirements of Art. 4 (Art. 8(1)). The methodology must be transparent, transferable, comparable, reproducible and easy to use (Art. 8(2)). Among other things, it covers the frequency of monitoring, the sampling, and arrangements for automatic and manual testing and for usability testing (Art. 8(3)(a) and (f)).

The methodology itself is set out in Implementing Decision (EU) 2018/1524. The first monitoring period for websites ran from 1 January 2020 to 22 December 2021, and the one for mobile applications from 23 June 2021 to 22 December 2021; since then, monitoring has taken place annually, each time from 1 January to 22 December (Art. 3(2), (3) and (5)). It is based on the standards and technical specifications referred to in Art. 6 of the Directive (Art. 4(1) of the Implementing Decision). For website content, Art. 6(3) of the Directive names EN 301 549 V1.1.2 (2015-04) for cases in which no references of harmonised standards have been published. The methods add nothing to these requirements and are independent of specific testing tools, operating systems, browsers or assistive technologies (Annex I point 1.1).

Reporting is less frequent than monitoring: by 23 December 2021 at the latest and every three years thereafter, Member States report the outcome to the Commission, including the measurement data and the use of the enforcement procedure (Art. 8(4) Directive (EU) 2016/2102). The report is published in an accessible format (Art. 11 of the Implementing Decision).

Simplified and in-depth monitoring: what is the difference?

In-depth monitoring establishes whether a website or app meets all requirements; simplified monitoring only looks for non-compliance with a subset of the requirements (Art. 5(a) and (b) Implementing Decision (EU) 2018/1524). The simplified method is intended for websites (Annex I point 1.3.1); Member States should consider applying it to mobile applications (recital 6).

How do documents end up in the sample?

Documents are monitored via the website to which they belong: for each type of service and each main purpose, the in-depth method covers at least one relevant downloadable document, if available (Annex I point 3.2(e) Implementing Decision (EU) 2018/1524). The monitoring body may add further pages it considers relevant (point (f)). In addition, pages are selected at random, amounting to at least 10 % of the sample established under points (a) to (f) (point (g)). For federal bodies in Germany, § 2a(1) BITV 2.0 makes clear that integrated content such as documents forms part of websites; § 2a(2) provides the same for mobile applications.

The websites to be monitored are chosen according to fixed criteria:

• a varied, representative and geographically balanced selection (Annex I point 2.2.1),
• all levels of administration, where they exist: state, regional, local and other bodies governed by public law (point 2.2.2),
• a variety of services, for example social protection, health, education, transport (point 2.2.3),
• consultation of national stakeholders, in particular organisations representing persons with disabilities (point 2.2.4),
• from the second period onwards, where the number of websites and apps allows, at least 10 % of those monitored in the previous period and at least 50 % that were not monitored in the preceding reporting period (point 2.4).

According to recital 5, the selection could be risk-based and could take into account, among other things, feedback received through the feedback mechanism and previous monitoring results.

Who carries out monitoring in Germany and Austria?

Germany

For federal bodies, monitoring is carried out by the federal monitoring body for the accessibility of information technology (Überwachungsstelle des Bundes für Barrierefreiheit von Informationstechnik), which is located at the Federal Accessibility Agency (Bundesfachstelle für Barrierefreiheit) (§ 13(3) BGG); the Federal Accessibility Agency is established at Deutsche Rentenversicherung Knappschaft-Bahn-See (§ 13(1) BGG). The monitoring body monitors periodically, advises the bodies on the basis of the test results, evaluates the reports of the supreme federal authorities and the Länder and prepares the report to the Commission (§ 13(3), second sentence, nos. 1 to 4 BGG).

The procedure follows Art. 1 to 7 and Annex I of the Implementing Decision (§ 8(1) BITV 2.0). The monitoring body records compliance with the requirements under Art. 6 of Directive (EU) 2016/2102 separately from the supplementary requirements of § 12a BGG and BITV 2.0, and may test usability (§ 8(2)). It may carry out ad hoc checks and repeat checks (§ 8(3)). It prepares the report to the Commission in accordance with Art. 8 to 11 and Annex II of the Implementing Decision (§ 9(1) BITV 2.0). The Länder report to it every three years, for the first time by 30 June 2021, in particular on the results of their own monitoring (§ 12c(2) BGG). The Berlin figures below come from the monitoring body for digital accessibility of the Land Berlin; this article does not cover the German state laws (Landesgesetze).

Austria

The WZG covers websites and apps of the federal government and federal-related bodies (§ 2(1) nos. 1 and 2 WZG). These must be monitored on a recurring basis, and every third year a report incorporating the reports of the Länder must be submitted to the Commission; the legal entities concerned must cooperate in the monitoring (§ 5(1) no. 1 WZG). The competent body is the FFG (Austrian Research Promotion Agency), as long as no regulation designates another body (§ 5(2)). It coordinates the checks with the Länder (§ 5(3)), examines complaints and issues recommendations for action where complaints are justified (§ 5(1) no. 2), and publishes the monitoring methodology (§ 5(1) no. 5). The Länder and municipalities are subject to the state laws (Landesgesetze); according to the FFG, the monitoring bodies of the federal states check annually on the basis of nine state laws. Details of the state laws examined are set out in the article Accessibility in Austrian municipalities.

What do the Berlin monitoring results for 2024 show?

In 2024, Berlin monitored a total of 117 websites and apps of public bodies of the Land and found an average compliance rate of 49.4 %; of the PDF documents, 7 out of 114 passed the automated test with the PDF Accessibility Checker (PAC), that is 6.1 %. Of these websites and apps, 108 were monitored with the simplified method, 4 with the in-depth method, and 5 were mobile applications; 12 were monitored again. Berlin calculates the compliance rate from the applicable requirements and the requirements met.

The requirements refer to clauses 9 (web) and 11 (software, except 11.5 to 11.8) of EN 301 549, which according to Berlin corresponds to the criteria of WCAG conformance level AA. Among the weakest results in 2024 were 1.3.1 Info and Relationships at 4.3 % (5 of 117) and 1.1.1 Non-text Content at 12.8 % (15 of 117). Berlin concludes that, on average, the websites and apps monitored are still clearly short of the legal requirements.

What does the FFG interim report for 2025 show?

None of the 23 websites given an in-depth check in 2025 met all requirements; all were partially compliant, which according to the FFG means that at least half of the roughly 92 criteria checked per website were met. The figures include the checks carried out by the FFG for the federal level and by the monitoring bodies of the federal states; the federal level accounted for 11 of the 23 in-depth and 174 of the 349 simplified website checks. The checks ran from 17 February to 16 October 2025 and comprised 23 in-depth website checks (around 440 subpages), 349 simplified website checks (around 5,400 subpages) and 16 in-depth app checks (around 200 screens).

Two findings are relevant to documents. WCAG criterion 1.3.1 Info and Relationships failed on every website given an in-depth check; among the most frequent errors under this criterion, the FFG lists untagged or incorrectly tagged PDF documents, meaning PDFs with no tag structure or an inadequate one, in the extreme case scans without any structural information. Under criterion 2.1.1 Keyboard, it lists links and form elements in PDF documents that cannot be reached with the keyboard. In 52.5 % of the statements, a disproportionate burden was claimed: often for missing audio description of videos hosted on YouTube, and several times with the statement that PDF documents could not be published accessibly as standard. When such a claim holds up is covered in the article Disproportionate burden.

Why is an automated test not enough?

An automated test finds only some of the barriers. The Implementing Decision itself describes the simplified method as a procedure that detects non-compliance with a specific subset of the requirements (Annex I point 1.3.1). The FFG states that simplified checks do not reveal all accessibility issues and cannot establish whether a website meets the criteria in full; only an in-depth manual check can do that.

For PDFs, the FAQ of the PAC publisher says the same: not all requirements of PDF/UA and WCAG can be checked by software. An additional check by a human is always necessary, for example to judge whether the semantic markup is appropriate. According to the FAQ, the result is not legally binding but an indicator; green ticks on the PDF/UA and WCAG tabs only mean that all requirements that can be checked automatically are met. PAC currently processes PDF 2.0 only on a best-effort basis; PDF 1.7 is fully supported.

The Berlin rate of 6.1 % therefore describes how many documents passed the automated PAC test. It does not say whether these documents also met the requirements that only humans can check; the remaining 107 of 114 documents already failed the machine check. How the standards fit together is explained in the article PDF/UA, WCAG 2.1 and EN 301 549.

What can a public body take from the findings?

The findings show where checks regularly fail: document structure, text alternatives and the accessibility statement. Practical steps can be derived from this; they do not replace legal advice in an individual case:

1. Sort documents by service. In-depth monitoring takes at least one document per type of service and main purpose (Annex I point 3.2(e)). Prioritise forms, information sheets and application forms that belong to a service.
2. Check structure before appearance. A tagged PDF with headings, lists and tables addresses one of the frequent errors the FFG lists under 1.3.1. Replace scanned documents without structure with versions generated from the source file; the checklist for Word and PowerPoint shows the steps.
3. Test by machine and by a human. Document both checks, because according to the PAC FAQ an automated report is only an indicator.
4. Keep your own assessment reports up to date. In in-depth monitoring, the monitoring body may use the body's own assessment results in whole or in part if the latest detailed report is submitted, the assessment is no more than three years old and was carried out in accordance with Annex I points 1.2.1 to 1.2.4 and point 3, and the monitoring body considers the report admissible (Annex I point 1.2.5).
5. Date and complete the statement. In the 2025 Austrian monitoring, 69 % of the statements recorded bore no current date (FFG); in Berlin in 2024, the statement was formally correct for 15 of 117 websites and apps monitored (12.8 %). For federal bodies in Germany, the statement must be updated annually and whenever there is a significant change (§ 7(6) BITV 2.0). The mandatory content is listed in the article on the accessibility statement.
6. Use the feedback. If monitoring identifies deficiencies, the body receives data on them within a reasonable period, in a format that helps to correct them (Art. 7 Implementing Decision).

An overview of obligations and deadlines is given in the guide to accessible PDFs. If you would rather not rework your document stock yourself, you can have Compliance-Kit make your documents accessible.