Harbor County Schools: concept study
Concept project. Harbor County Schools is a fictional school district. This is a Thirdhall design and engineering demonstration, not a client engagement. The “before” data comes from real public websites; the “after” is the concept, measured the same way. View the concept
Before: what sites like this look like today
We scanned Rhode Island's 36 school district websites for WCAG 2.1 AA failures on the homepage and a key task page.
- 69%had a serious or critical accessibility failure (25 of 36)
- 30link to no accessibility statement
- 7rely on an overlay widget instead of fixing code
- 68average benchmark score of two large Rhode Island school district websites
The problem we designed against
Families come to a district website with a handful of urgent questions: is school open, which school is mine, how do I enroll, what's for lunch. District sites often bury those under news and board material, publish menus and calendars as PDFs, and rarely say where to get help in another language.
What the concept does
- Today's open or closed status at the very top, then the six questions families ask most.
- An address-based school finder that returns all three assigned schools and the bus stop.
- Enrollment as four numbered steps, with documents listed and language help named up front.
- Menus and calendars as web pages, readable by screen readers and translation tools.
- A Spanish homepage, linked with proper language markup from every page.
Before and after, same benchmark
| Area | Typical real site (avg of 2) | Concept |
|---|---|---|
| Accessibility | 52 | 100 |
| Documents | 73 | n/a |
| Mobile performance | 52 | 99 |
| Security & maintenance | 55 | 80 |
| Content health | 100 | 100 |
| Language access | 85 | 100 |
| Privacy | 90 | 100 |
| Overall | 68 (D) | 96 (A) |
What this shows, and what it doesn't
- It shows the design patterns and build quality we deliver, measured with the same tools we use on real sites.
- It does not show results with real residents, staff, or traffic. A concept has no users, no legacy content, and no integrations, which makes high scores easier. Those are where real projects get hard.
- Automated checks find a subset of accessibility problems. Real projects add manual testing with assistive technology.
How we'd measure a real engagement
- Benchmark before launch and after, on the same rubric.
- Task completion for the top errands, not page views.
- Calls and emails about questions the site should answer.
- Share of documents that are accessible, and time to publish one.