Haven Valley Unified

Every file was tagged. Almost none of them worked.

A Bay Area school district had done what many districts do: it ran its PDFs through an automatic tagging tool. We checked whether the tags actually work for a student or parent who uses a screen reader.

Client
A Bay Area school district
Type
Public document review
Scope
22 recent public PDFs
Result
21 of 22 had barriers

Context and goals

Districts post board packets, budgets, and family notices every week. Parents who are blind or have low vision need to read them like anyone else.

We reviewed the district’s recent public PDFs to see whether its tagging was doing its job.

What success looks like

  • A screen reader announces headings, so a parent can jump to the section they need.
  • Every meaningful image has a short description, and decorative images are marked so they are skipped.
  • Tables read in order, with their headers announced.

What we measured

We checked 22 PDFs the district posted in 2025 and 2026 for tags, headings, image descriptions, language, and title, then checked the worst files by hand.

How many of the 22 files passed each check

  • Tagged22 of 22
  • Named its language22 of 22
  • Had a title22 of 22
  • Had headings3 of 22
  • Described every image1 of 22

The tags were there. The parts a parent actually hears were missing.

A sample of the files we checked
DocumentPagesTaggedHeadingsImages described
Weekly newsletter, October 20254YesNone0 of 3
Weekly newsletter, August 20255Yes30 of 3
Board newsletter, June 20264YesNone0 of 1
Weekly newsletter, October 20264YesNone1 of 2
Weekly newsletter, November 20263YesNone0 of 2
Public notice, November 20251YesNoneNo images

What we found

  • The good news: all 22 files were tagged, which means the district cares and already has a process.
  • 21 of the 22 had images with no description, so a screen reader says nothing or reads a file name.
  • 19 of the 22 had no headings, so a parent hears one long wall of text with no way to jump ahead.
  • The pattern points to automatic tagging that nobody reviewed afterward.

A four-page weekly family newsletter, posted as a PDF

One file, marked up

  1. The newsletter’s title is tagged as plain text, not a heading.
  2. Every photo has no description, so a screen reader skips it or says “graphic.”
  3. Section titles are plain text too, so a parent can’t jump to the part they need.

After: the tags we add

  • <H1> Weekly Community Update
  • <Figure alt="Students at a fall science night">
  • <H2> Upcoming events
  • <L> Lists tagged as lists
  • <Artifact> Decorative border, skipped

Hear the difference

A parent who is blind opens the weekly newsletter and looks for upcoming events.

Before

What a screen reader says

  1. “Weekly Community Update.”
  2. “Graphic.”
  3. Then every word, top to bottom. There are no headings to jump to, so the parent listens from the start every week.

After

What a screen reader says

  1. “Heading level 1, Weekly Community Update.”
  2. “Graphic, Students at a fall science night.”
  3. “Heading level 2, Upcoming events.” The parent jumps straight there.

The path to good

  • Review each tagged file by hand: add headings, describe meaningful images, mark decorative ones, and check tables.
  • Train the staff who make these files to build headings and descriptions in Word, so the export starts out right.
  • Use a monthly retainer so every new file is reviewed before it goes up.

Follow-up

This was a review of public files, not a paid engagement, and the district’s name is kept private.

Take it with you

Download the work.

Both files are tagged and tested for screen readers.

See more

See how other organizations are doing.

See all case studies

Want to see it on your files?

Get a free accessibility check.

Start free