How to Conduct an Accessibility Audit 

Accessibility audits are often treated as the point at which accessibility becomes visible. 

In many organisational contexts, they function as a signal that accessibility has been addressed because they produce reports, identify failures, and offer reassurance that requirements have been acknowledged.  

That signal is not meaningless, but it is frequently misinterpreted. 

The core framework laid out in Ensuring Digital Accessibility in Learning Design demonstrates that technical standards and audits provide infrastructure, not guarantees. They stabilise expectations and enable accountability, but they do not determine whether a learning design is valid in use or really how accessible it is (something that WCAG 3.0 hopes to address). What they show is partial, shaped by the conditions under which they operate, and limited in what can reasonably be claimed from them. 

What an audit makes visible 

Accessibility audits check against criteria, surface technical failures, and make certain forms of inaccessibility legible. In doing so, they create a shared language for identifying and discussing accessibility issues. This is where their value lies. 

At the same time, the act of defining what is checked shapes what becomes visible. Audit criteria inherently privilege certain forms of performance that can be tested reliably, rendering alternative barriers invisible. What falls outside those parameters is not necessarily absent, but it becomes easier to disregard. 

Accessibility audits operate within defined parameters, so this is not a flaw in the audit, it is a consequence of its design. 

What remains out of view 

A design can meet accessibility standards while still failing under real conditions. 

An accessibility audit confirms that components are present and technically compliant. It does not determine whether the design remains coherent once those components are used. So, barriers arising from changes to pacing, sequencing, or the interaction between modalities, etc not being accounted for in the design are less likely to be captured through standard audit criteria. 

This creates a familiar tension. 

A learning experience may: 

• include captions 

• meet contrast thresholds 

• provide structured navigation 

and still require learners to reconstruct meaning, manage fragmented attention, or compensate for assumptions embedded in the design. 

In these conditions, accessibility is present, but not equivalent. The audit has identified compliance; it has not demonstrated design validity. Access to elements is not the same as access to meaning, and technical participation is not, on its own, evidence of an equivalent learning experience. 

The problem with late-stage assurance 

The audit becomes a mechanism of assurance when used as a final checkpoint and is structurally positioned too late to affect meaningful technical changes let alone design changes.  

In practice, by the point audits often occur, the structure of the design is already fixed. Decisions about pacing, modality, interaction, and assessment have been made so the audit evaluates what already exists rather than guide or evaluate decisions within the design process. It confirms that requirements have been met without interrogating whether those requirements were sufficient, or whether the design rests on assumptions that do not hold under variation. 

Accessibility becomes something that has been checked, rather than something that has shaped design. It functions symbolically, present in documentation but only partially realised in practice. 

Audit as part of design 

Accessibility should not be introduced at the end as a point to check off. It should be established through earlier decisions about structure, sequencing, interaction, and assessment but by the time an audit is conducted, many of those decisions are already embedded. 

This is why audits tend to surface technical gaps more reliably than structural ones. They operate within the constraints of existing design. The limitations of audits do not make them irrelevant however; they should make their positioning more precise. 

If accessibility is treated as a continuous design responsibility, the role of the audit changes. It becomes a way of identifying gaps and testing consistency throughout the design process. Not a substitute for design literacy or judgement but a tool to be used to guide decisions. 

Judgement under constraint 

An accessibility audit can show that criteria have been met and that specific forms of inaccessibility have been addressed. It provides evidence under particular conditions, shaped by particular standards.  

That evidence supports only limited forms of inference about accessibility in practice. It cannot establish that a learning design is inclusive, coherent, or meaningful in use. The presence of captions or structure does not, on its own, demonstrate that learners can interpret or engage with the design as intended. Recognising this does not weaken the role of the audit. It constrains what can be claimed from it. 

Once audits are recognised as insufficient, responsibility shifts so that decisions that might otherwise be deferred to standards or checklists return to designers and teams. Questions that appear resolved through compliance re-emerge as matters of judgement. 

Different forms of access do not always align but these are not problems that audits can resolve, they are conditions within which they operate.  

Designs that meet technical standards may still privilege certain forms of engagement. Designs that broaden access may introduce complexity or challenge institutional expectations of consistency. 

This introduces a tension that rarely sits well with traditional institutional metrics.  Leadership frequently defaults to seeing accessibility checklists as a bureaucratic compliance mechanism to mitigate legal risk as quickly and cheaply as possible while churning out faster, cheaper courses.  

This view treats accessibility as a feature to be audited rather than a condition under which design operates. But for those responsible for learning architecture, standards are a floor, not a ceiling.  

Chasing fast, cheap compliance obscures a deeper systemic failure: a design that depends on idealised conditions to lower immediate production costs simply defers those costs onto the learners. When our designs leave learners to navigate fragmented experiences, institutional consistency becomes a proxy for systemic exclusion. Shifting this paradigm requires recognizing that true efficiency isn’t found in a passed audit report; it is found in creating learning architectures that hold under real conditions, even when foundational assumptions fail. 

Closing position 

Accessibility audits reveal some failures, obscure others, and provide a partial account of how accessibility is enacted. Treated as endpoints, they risk turning accessibility into a matter of compliance and documentation. But accessibility audits should not sit outside design. They should sit within it, shaped by its assumptions and constraints. Treated as part of a wider design responsibility, they support more careful and accountable practice. 

The distinction lies in what is being claimed. 

Accessibility cannot be verified at a single point in time. It emerges from decisions that either hold, or fail, under real conditions. Audits can make some of those decisions visible, but they cannot make them sufficient.