You Have 5,000 Legacy PDFs. Which Ones Do You Actually Have to Fix?

Somebody in your organization has now run the report, and the number came back in the thousands. Board minutes going back a decade. Permit applications. Budget appendices. Notices nobody has opened since the year they were posted. The instinctive reaction is that all of it has to be remediated, that the cost is unthinkable, and that the project is therefore impossible to start.

It is not all in scope. The rule that created this obligation also drew lines around it, and those lines are the most useful triage tool you have. Before you price a single file, you should know which shelf each one sits on.

Does the rule really exempt anything?

Yes. The ADA Title II web rule contains five exceptions at 28 CFR 35.201, and two of them are aimed squarely at the pile you are staring at.

Preexisting conventional electronic documents. “Conventional electronic documents” is the rule’s term for PDFs, word processor files, presentations and spreadsheets. If one existed before your compliance date, it is excepted — unless it is currently used to apply for, gain access to, or participate in your programs, services or activities. That last clause is the whole ballgame, and we will come back to it.

Archived web content. Content is archived only if it meets every one of four conditions: it was created before your compliance date (or reproduces paper or physical media from before it), it is kept exclusively for reference, research or recordkeeping, it has not been altered or updated since archiving, and it lives in a dedicated area clearly identified as an archive. All four. A file that is merely old is not archived.

The other three exceptions cover content posted by genuinely independent third parties, individualized password-protected documents about a specific person’s case or account, and social media posts made before your compliance date.

Which compliance date is yours depends on what kind of entity you are, and those dates are listed in our ADA Title II deadline post — the one place we keep them, so a future change is corrected once.

The question that sorts the whole pile

Every exception above turns on one distinction, and it is not age. It is function.

Ask of each document: can a member of the public do something with this? Apply for a permit. Claim a benefit. Register a child. Contest a bill. Understand a rule that binds them. Attend a hearing. If yes, it is in scope regardless of when it was made, because it is currently used to participate in your programs.

If the honest answer is that the document exists so a researcher or auditor can find it later, and nothing about it is live, it is a candidate for the archive exception — provided you actually treat it like an archive.

That is the catch most entities miss. The archive exception is not a status you assert about a file; it is a condition you have to maintain. Content scattered across your live site with a 2014 date is not archived. Content in a clearly labelled archive area, untouched since it went in, is. If you want the exception, you have to do the work of building the archive.

A four-tier triage

Sort the backlog into four buckets. Most agencies find the top two are a small fraction of the total.

Tier 1 — Transactional and time-critical. Anything the public fills in, submits, or must act on by a date. Applications, claim forms, tax notices, hearing notices, emergency and public-safety information. These are in scope, they carry the highest real-world harm when inaccessible, and they should be remediated first even if you had no deadline pressure at all. A resident who cannot read a boil-water notice is a failure long before it is a finding.

Tier 2 — Currently governing. Documents that state a rule, a rate, a right or a schedule that is still in force. Fee schedules, current policies, service maps, adopted ordinances, benefit eligibility criteria. In scope, high traffic, and the ones staff link to daily.

Tier 3 — Current-year informational. Newsletters, reports, presentations, this year’s board packets. In scope going forward. The important move here is not remediation, it is turning off the tap — see below.

Tier 4 — Genuine reference. Historical minutes, superseded budgets, closed-project records. Candidates for the archive exception if you will genuinely build and label an archive, and if you will leave them alone once they are in it.

Then run one filter across all four tiers: pull your analytics. Sort every PDF by downloads over the last twelve months. The distribution is almost always lopsided — a small set of documents accounts for most of the traffic, and a long tail has never been opened. A file in Tier 3 with heavy traffic outranks a file in Tier 2 that nobody has touched. Real usage beats intuition about importance every time.

Stop the tap before you bail the boat

Here is the failure that turns a finite project into a permanent one: an agency remediates two thousand documents over eighteen months while continuing to publish new inaccessible ones the entire time. The backlog does not shrink. It relocates.

Fix the intake before, or at least alongside, the backlog:

• Train the handful of people who actually produce documents. In most agencies it is a small group — clerks, communications staff, a few program managers. Structure applied in Word before export is far cheaper than tags repaired in a PDF afterwards, which is the case we make in accessible Word, PowerPoint and Excel documents.
• Make accessibility a publication gate, not a later cleanup pass.
• Put the requirement into contracts. If a consultant delivers your master plan as an untagged PDF, you have bought a remediation job you did not budget for. The same clause logic applies as in writing accessibility into an RFP or scope of work.
• Ask whether the document needs to be a PDF at all. A fee schedule as an accessible HTML page is easier to maintain, easier to translate, and easier to keep compliant than the same content as a file.

Where translated documents sit in the queue

If you publish in more than one language, your Spanish, Vietnamese, Arabic or Somali files belong in the same tier as their English counterparts, not in a lower one. A Tier 1 application form is Tier 1 in every language you offer it in.

This is the point where two obligations meet. Title VI of the Civil Rights Act requires meaningful access for people with limited English proficiency. The accessibility rules require the documents to work with assistive technology. A translated form posted as an untagged PDF satisfies neither — and multilingual files carry failure modes English ones do not, including per-passage language tagging and reading-order corruption in right-to-left scripts. We cover the mechanics in translated documents and WCAG language of parts.

If you are budgeting a translation refresh and a remediation project separately, combine them. Remediating a document and then translating it means paying to fix structure twice.

What to do this month

Do not start by remediating. Start by counting.

Inventory every document with its URL, date, owning department, language, and twelve-month download count. Tier each one by function, not age. Decide honestly whether you will build a real archive area, because that decision determines whether Tier 4 is work or not. Then price Tier 1 — and only Tier 1 — and get it moving.

A backlog you have measured is a project. A backlog you have only estimated is an excuse to keep postponing it.

Ready to scope yours? Send us your document inventory — even a rough spreadsheet of URLs with page counts and languages — and request a quote. We will come back with a tiered remediation plan and a per-tier estimate. If you want to see what you are dealing with first, run a representative file through our PDF accessibility checker, or read how the remediation work itself is done on our document accessibility page.