Reporting Up, Your Maturity Roadmap, and the Budget Ask
Trust note
This article was written by a person and reviewed against accessibility.chat editorial standards. Treat it as research and education rather than legal advice. We prioritize primary sources and correct material errors.

A metric without a price is information. A metric with a price is a budget request.
This is the last piece in this series, so let me close with the part that decides whether everything before it survives past year one: getting your compliance committee to understand what they're looking at, and what it costs to improve it.
Five or six numbers, each with a price
Top-patient-task conformance, tested with an actual keyboard and screen reader against WCAG 2.1 AA (opens in new window) — not a scanner count. Backlog size and age. The share of active contracts carrying accessibility clauses. Complaint volume and resolution time. Training completion by role.
Never report one composite "accessibility score" — it hides exactly what a committee needs to see. And pair every number with a price: "backlog: 200 items" is information; "backlog: 200 items, clearing the top quarter costs sixty hours" is a decision someone can actually make.
Score yourself honestly, on a ladder that doesn't flatter you
Eight dimensions — policy and resourcing, inventory and testing, procurement, publishing workflow, training, feedback and complaints, exceptions and records, measurement — each on a four-rung ladder: Ad Hoc (things happen when someone complains, no owner, no record), Planned (a policy exists but no hours or money are attached — the rung where good intentions go to die politely), Resourced (hours in writing, gates actually running), Measured (the numbers above, tracked year over year with prices attached).
"Planned" is the rung that kills small programs, because it feels like progress while nothing actually moves. That's why the ask below leads with hours, not a headline number.
Three funding levels, presented together
A single number invites a no. A menu invites a choice.
Level zero — hours only. A fraction of one person's time, in writing. Buys the free things: overdue 1557 basics confirmed, the intake log, procurement clauses, the publication gate, honest reporting. Removes most avoidable exposure. Doesn't buy backlog movement or testing depth — say so plainly.
Level one — a modest line. Hours plus a small annual budget: paid structured testing with disabled patients twice a year, role-based training time, remediation for the top of the triage queue, one manual test cycle of top patient tasks. For most organizations this size, the honest target for year one.
Level two — funded. A real line item and named fractional roles: professional audit support on a cycle, systematic backlog work, tooling that enforces the gates instead of depending on goodwill. Ask for this level with a year of honest numbers already in hand — a funded request built on real data is the easiest conversation in this entire series.
Whatever's decided, get it in writing, including a decline. A documented no still answers the question that matters later: did the coordinator ask?
One complaint, the whole machine
A patient with low vision calls patient relations: his discharge instructions render as a scanned image in the portal, and he needs to change a dressing tonight. Logged within the hour, acknowledged on the call. The discharging unit's nurse calls him, walks through the instructions verbally, confirms teach-back, and emails a plain-text version through the portal's message function — done by end of shift, documented as a patient-safety event because that's what it was.
The next day: reproduce and classify. Every patient on that unit gets the same scanned image — an artifact of an old system migration — so this is systemic, not isolated. Two tracks open: the workflow fix (the EHR already exports instructions digitally; the scan step gets deleted, an IT ticket, not a project) and the vendor track (the portal's document rendering goes into the vendor review file, and the ACR gets requested at the next business review). The BAA question settles the backlog decision automatically — no outside vendor touches these documents.
A follow-up call closes the loop with the patient. The log entry gets its resolution date. Total cost: about six staff hours. The uncorrected version of this story is a readmission, then a complaint, then an investigator (opens in new window) reading a chart that shows the organization knew.
The part worth remembering
Somewhere in your patient population is someone who couldn't read a form, couldn't operate a portal, couldn't complete a telehealth visit on equal footing — and didn't file a complaint about it. They just stopped trying. Every chapter in this series exists to serve the one fact that you're the person in the building whose job is to notice that. Go build the boring parts. They're the ones that actually hold.
None of this series is legal advice — it's a field map, meant to be checked against your own counsel and your organization's specific facts. If you want a fast first read on a page before you bring it to your compliance committee, ask Luke. Our accessibility statement and methodology show this same discipline applied to our own site.
Previous: building the record · Start over: you're the 1557 coordinator, the law behind the job
Sources: Web Content Accessibility Guidelines (WCAG) 2.1 (opens in new window) · HHS Guidance on Section 1557 (opens in new window) · HHS Final Rule — Nondiscrimination in Health Programs and Activities (opens in new window) · U.S. Access Board — Medical Diagnostic Equipment Standards (opens in new window)
About Jeff Fryer
Jeff Fryer spent years working kitchens before moving into ADA compliance work for local government. He writes from that experience -- direct, plainspoken, allergic to compliance theater. Contributing writer at accessibility.chat.
Jeff Fryer is a person, not one of the AI analyst lenses this site also publishes under. A named human is accountable for this article.
Specialization: Local government ADA compliance, contributed from direct field experience
Authorship and Editorial Process
Jeff Fryer wrote this article. AI was not used to draft it. It went through the same editorial checks as everything else published here.