DAP Quality Benchmarking Tool: Automated accessibility benchmarking for the most-visited government web pages

The problem

Government websites can contain thousands of pages. Accessibility problems can exist on any of them, but the pages that receive the most visitors are where barriers affect the greatest number of people. A broken form on a page that ten people visit each month is a different priority from the same problem on a page that ten thousand people use every day.

Monitoring and improving the quality of government digital services, especially the most-visited sites, requires regular and consistent data. A single accessibility scan gives a snapshot, but it does not show whether things are getting better or worse over time. A team might fix a set of problems one month only for new ones to appear the next, and without regular, comparable data it is difficult to identify this pattern. For organisations that want to track whether their accessibility work is making a difference, or that need to compare performance across multiple websites, periodic checks are not enough.

The solution

The US federal government already collects data on how the public uses its websites through the Digital Analytics Program (DAP), a shared web analytics service run by the General Services Administration. DAP tracks which pages people visit, how often, and how they interact with government websites. Federal agencies are encouraged to participate.

DAP tells agencies which sites get the most traffic. What it does not do is check whether those pages are accessible to people with disabilities or meeting quality standards.

The DAP Quality Benchmarking Tool was built to connect those two things. It takes the visitor data that DAP already collects and uses it to check the most popular pages for accessibility and quality. The reasoning is straightforward: problems on the most-visited pages affect the most people, so those are the pages that should be monitored first.

Each day, the tool automatically selects the top pages by visitor traffic and checks them using two established testing tools. The first is Lighthouse, built by Google, which assesses a web page's performance, accessibility, and overall quality. The second is Axe, which provides specifics on the accessibility issues. Both tools work by checking pages against recognised standards using fixed rules.

For each accessibility issue the tool finds, it does something that most scanning tools do not: it estimates how many people are likely to be affected. Using disability prevalence data from the US Census Bureau, the tool calculates roughly how many of a page's visitors may be unable to use it because of a specific barrier. For example, if a page that receives millions of daily visitors has a problem, we can get a good estimate on the number of people affected. The tool estimates how many of those visitors fall into that group based on national population data. This turns what might otherwise be an abstract technical finding into a number that helps teams understand who is being affected and how many, making it easier to explain why a particular fix should be prioritised.

The results are published as reports dated to each day the tool runs, with a history that allows teams to look back and compare. This means teams can see whether accessibility on their most-visited pages is improving or declining over time, and can compare performance across different websites. For teams reporting to leadership on whether accessibility investment is producing results, the trend data provides evidence that a single one-off scan cannot.