The core procedure provides five sequential steps, and each step has its own sub-steps.
- Define the evaluation scope:
After a brief exploration of the site and conversations with the responsible stakeholders, we define:
- The scope of the website, or which pages are going to be evaluated.
- The conformance target (“A”, “AA”, or “AAA”)
- An accessibility support baseline, or the minimum set of combinations of user agents (operating systems, web browsers, assistive technologies…) that the website is expected to work with.
- Other optional additional evaluation requirements
- Explore the target website:
Here we identify relevant pages of the site, and maybe we change the evaluation scope previously defined. This exploration includes the identification of:
- Common web pages
- Essential functionalities
- Web page types and states
- Web technologies relied upon
- Other relevant web pages
- Select a representative sample:
Depending on the size, the age, the complexity and consistency of the site, the adherence to development processes, the required level of confidence and the availability of prior evaluation findings, we will include:
- A structured sample
- A randomly selected sample
- Complete processes
- Audit the selected sample:
This is the moment when we check all of the web pages and web page states selected in the third step with the Techniques procedures. Precisely we check:
- All initial web pages
- All complete processes
- And compare structured and random samples
- Report the Evaluation Findings:
Finally, it is time to document if the pages have passed or not the tests, where we provide
- The outcomes of each step, with the name of the evaluator, the date, the scope, the exploration, the representative sample and the sample audited.
- Optionally, the record of the evaluation specifications, that is, an archive of the web pages and states, evaluation tools, browsers…
- Optionally, an evaluation statement describing the outcomes of the conformance evaluation (not a conformance claim, but similar).
- Optionally, an aggregated score, to understand the evolution in quantitative data. Although the WCAG 2 does not provide scoring metrics, the W3C is researching on it.
- Optionally, machine-readable reports in the Evaluation and Report Language (EARL)