accessrank.ai · Accessibility Evidence File · Sample
Accessrank, Inc. · info@accessrank.ai · Los Angeles, CA

Sample record. This is a shortened example of the Accessibility Evidence File Accessrank prepares for each client. It is based on a real engagement with every identifying detail removed: the store name, website, people, products and contact details are invented, and dates are left out. A real record dates every entry. Remediation can take anywhere from a few days to a few months, depending on the site.

Accessibility Evidence File

Example Store · store.example · Client packet · Version 1.0

Accessrank, Inc. · Access & Discovery Solutions · accessrank.ai

Contents

SectionWhat it answers
Cover memoWhat this record is, key dates, and how to use it
Section 1 — Baseline auditWhat the store knew, and when
Section 2 — Remediation logWhat has been fixed, and when
Section 3 — Conformance summaryWhere the store stands now, and what was tested
Section 4 — Public accessibility statementWhether the program is public, and the inquiry log
Section 5 — Ongoing monitoring recordWhether the program is continuous: every scan, dated
Section 6 — Manual and assistive-technology testingWhether a person has verified the work
Section 7 — Program engagement and scopeWhat the store has committed to, and known limits
Language guidelines and disclaimerWhat to say, what not to say, and the limits of this record
Exhibits and attestation formsSupporting documents and signable forms

Status at a glance

Protection levelDeter: every finding an automated scan can see is fixed and monitored. Verify: human testing in progress (Section 6)
Lighthouse accessibility88 at baseline; 93 after the first round of fixes; 100 after the second, and 100 again at the latest re-test
Automated violations0 WCAG 2.2 A/AA violations on an 11-page re-test of the live site
Remediation log42 fixes and 3 approved changes. 40 fixes closed and passed on re-test; 2 closed in code and waiting for confirmation by a person (Section 6)
Accessibility statementApproved by the store owner and published, linked from every page footer (Section 4)
Still openScreen reader and phone testing (Section 6); audio description for one brand video, awaiting the owner’s decision

Cover memo

↑ Contents

This record documents the website accessibility program for store.example, which began with a baseline audit before any fixes. It is maintained by Accessrank, Inc. and refreshed after each scan and on request. Every item is dated, and every finding is measured against WCAG 2.2 Level A and AA.

It is organized so a reader, such as the store’s attorney or insurer, can answer three questions quickly: what the store knew and when, what has been fixed and when, and what the store is doing on an ongoing basis. Section 1 establishes the starting point. Sections 2 and 3 show the fixes and where the store stands now. Sections 4 through 7 show whether the program is public, continuous, checked by a person, and committed.

StageBy
Baseline audit, before any fixesStore + Accessrank
Remediation, in rounds, each published and re-testedAccessrank, Inc.
Owner approvals of every visible changeStore owner
Accessibility statement approved and publishedStore owner; Accessrank
Automated re-test after each publishAccessrank, Inc.
Manual and assistive-technology testingAccessrank + testers (in progress)
Monitoring and certificatesAccessrank, Inc., quarterly and on request

Section 1

Baseline audit

↑ Contents

The baseline records the state of store.example before any fixes. It shows, with date stamps, that the store identified its own accessibility issues and acted on them.

ItemBaseline
AuditGoogle Lighthouse accessibility report of the home page, and a review of product and collection pages by Accessrank (12 products, 2 collections)
ToolingGoogle Lighthouse; Accessrank automated scan of the live site in a real browser against WCAG 2.2 A/AA
ScopeCustomer-facing pages and templates. Checkout and customer account pages are hosted by the store platform and are out of scope
Lighthouse accessibility score88 / 100
Findings4 failed Lighthouse checks on the home page (a pop-up with no name, focusable items inside hidden areas, low-contrast text, small touch targets), plus missing image descriptions found in the page review
AttestationSigned by Accessrank: Form A1, attached

Section 2

Remediation log

↑ Contents

Every fix below was made in the store’s theme code or content, not through an overlay or widget. Theme changes were tested first on a copy of the theme and then published. In a real record every entry carries the date it went live; dates are left out of this sample. Entries are never edited after the fact; a correction is a new row. The full log is shown: 42 fixes and 3 approved changes, newest first.

FindingWCAGPagesHow fixedStatus
Accessibility statement (approved by the store owner)—Site-widePublished as approved and linked from every page footerClosed
Live
Accessrank footer credit (approved by the store owner)—Site-wide footerAdded as shown in the approval packetClosed
Live
Who handles accessibility reports (approved by the store owner)—Accessibility statementAccessrank receives reports and replies for the store, with a copy to the owner, within 2 business daysClosed
Live
A sold-out product button was announced to screen readers as disabled, but pressing Space on it still fired a click. A control that says it is disabled but still activates is a name, role and value fault. Traced by instrumenting the browser’s click function on the live page: the call came from an inline script in the theme layout, not from the theme’s asset files, which is why a search of those files had not found it4.1.2ShopA guard in the theme’s accessibility script stops the Space key on the two sold-out controls only, and only while they are marked disabled. For the visitor the key does nothing and the page does not scroll; what changes is that no click is fired
Re-test: Published on its own, on a copy of the live theme. Before publishing, every layout, section, snippet and template file was compared with the live theme by checksum: only the two intended files differed. On the live site, Space on a sold-out control produced no click and focus stayed put; on an available product Space still adds to the cart
Closed
Passed re-test
Each product tile’s “View” link opens a product panel in place (the page goes from 339 to 533 visible elements), but the link carried no expanded state and no reference to the panel, so a screen reader user was not told that anything had opened, or where4.1.2ShopEach link now names the panel it controls and reports whether it is open, read from what the panel is actually showing rather than from the address bar. The links keep working as links if scripts do not run
Re-test: All 12 product links checked on the live site: collapsed on load, and opening two products leaves exactly those two reported as open. Published on its own; all 377 theme files compared before publishing, one changed
Closed
Passed re-test
All 28 quantity buttons on the shop grid were named only “Increase quantity” or “Decrease quantity.” Read as a list of controls, nothing told twelve products’ buttons apart2.4.6
best practice
ShopEach button is now named for its product; the four inside the product panel take the name of the product that is open
Re-test: All 24 tile buttons named for their product on the live site; the panel buttons fall back to a generic label when no product is showing
Closed
Passed re-test
Correction to the captions entry below. The first caption file was timed from silence detection over the film’s music bed rather than from measured speech, so its cues did not match the audio. Captions that are not synchronised do not satisfy 1.2.2, so that entry overstated the position until this correction1.2.2AboutRebuilt from measurement. A speech detector separated speech from music (18 segments, 173 seconds of speech); speech recognition then gave word-level timings for 468 words, aligned to the approved caption text (81% matched directly, the rest interpolated across short gaps). Cues were re-cut at sentence boundaries so no caption appears before it is spoken; the old file ran up to five seconds early
Re-test: On the live site: 43 cues, all parsed by the browser, captions showing. Checked against three timings the store owner read from the player clock (0:09, 1:03 and 3:02) against 9.30, 62.74 and 182.28 seconds in the new file. Two lines at most, none faster than 21 characters per second, and every word of the approved text kept
Closed
Passed re-test
The new text alternative for the video set white text on the About page’s light-blue gradient: 3.44:1 for dialogue, 3.23:1 for the summary, and 2.90:1 and 2.83:1 for the descriptions. The fix for one criterion had introduced a failure of another1.4.3AboutCovered by the flat page background described in the heading entry below
Re-test: Re-measured on the live page: summary and dialogue 9.46:1; notes and descriptions 7.35:1, against 4.5:1 required
Closed
Passed re-test
The text alternative’s scrolling box (1,466 px of scroll) contained nothing focusable, so a keyboard user could not reach it or scroll it, and everything below its fold was out of reach without a mouse2.1.1AboutThe box is now a named, focusable region with a visible focus outline
Re-test: On the live page it takes focus and scrolls by keyboard
Closed
Passed re-test
The heading above the video, white 40 px bold type, sat directly on the page’s light-blue gradient: 2.36:1, under the 3:1 allowed even for large text. Pre-existing, not caused by the video work1.4.3AboutThe gradient was replaced by one flat blue, the dark color the gradient already ended on, so it is the same brand color. No panel or bar behind any text. Limited to the About page
Re-test: White text now 9.46:1. A sweep of every visible text element on that background found no failures. Automated scan: 0 violations, 34 passes. The 30 items the scanner could not decide (text over the footer’s gradient) were measured by hand against the gradient’s own colors: all 8 text styles pass
Closed
Passed re-test
The brand video had captions but no text alternative, so a visitor who cannot see it had no account of what it shows1.2.3AboutA full text alternative under the player, opened and closed by keyboard: 21 descriptions of what is on screen interleaved with 19 passages of dialogue, covering all 35 shots, written after scene detection and a frame-by-frame review
Re-test: Live; closed by default, with a 46 px control. Speakers are not labeled line by line, and the record says why: who is on screen was verified from the footage, who is speaking was not
Closed
Passed re-test
The 3-minute brand video was embedded from a channel the store does not manage, so its only captions were the automatic ones: no punctuation, no speaker or sound cues, no checked accuracy, and turned off by default1.2.2, 1.2.3 (partial)AboutThe video is now served from the store’s own files as a native video with a caption track, shown by default. Re-encoded at 1920×1080, and the player corrected to a true 16:9 ratio, which also removed the black bars
Re-test: Live: 42 caption cues parsed and showing by default; automated scan 0 violations. Caption text came from speech recognition and was corrected by hand against the audio. Audio description (1.2.5) remains open and is listed as an open item
Closed
Passed re-test
Following an in-page link put the target under the sticky header, with all 161 px of it covered, including the main content the skip link jumps to2.4.11Site-wideScroll spacing tied to the theme’s own header height, so targets stop below the header at every screen width
Re-test: Target now lands at 161 px, level with the header’s bottom edge: 0 px covered, where it had been 161 px. Automated scan 0 violations
Closed
Passed re-test
The header account link was 0 px wide: invisible and impossible to click, but still reached by the Tab key and read aloud. It was the only route to the account area anywhere on the site2.5.8, 4.1.2Site-wide headerThe link was moved, not copied, into the top strip beside the cart and given a real size, so only one account link exists
Re-test: 73 × 28 px on desktop and 44 × 28 px on phones, against a 24 × 24 minimum; named “Log in,” and “Account” when signed in; 6.73:1 contrast; visible focus outline; 0 violations
Closed
Passed re-test
On phones, the shipping notice in the top strip ran underneath the cart and account icons, and the strip did not line up with the header below it1.4.10, 1.4.4
supporting
Site-wide headerThe strip now lines up with the header column, and the notice reserves room measured from the icons, switching to a shorter sentence only when the full one will not fit
Re-test: One line at every width from 2,155 px down to 320 px, and never reaching the icons
Closed
Passed re-test
The related-products row on product pages was held at near-zero opacity until scrolled into view, so its heading, 12 product names and prices measured 1.01:1, near-white on white1.4.3Product pagesHidden state corrected to fully hidden, respecting the visitor’s reduced-motion setting, with a fallback that shows the row if scripts never run
Re-test: 0 violations on the live product page measured from the top, where the previous theme returned 26; the fade-in still runs
Closed
Passed re-test
A large headline over a bright part of a photo was hard to read1.4.3ShopDark backing behind the headline only; photo and products not darkened
Re-test: Lowest point 4.1:1 (large text needs 3:1)
Closed
Passed re-test
A heading and first paragraph over a photo were hard to read1.4.3Content pageDark backing behind the heading and paragraph
Re-test: Heading 8.8:1 and paragraph 7.9:1 at worst
Closed
Passed re-test
An embedded video had no name4.1.2AboutDescriptive title added to the video frame
Re-test: Title present on the live page; 0 violations
Closed
Passed re-test
The store locator’s main title was plain text, not a heading1.3.1Store locatorMarked as a level-2 heading
Re-test: Appears in the page’s heading outline
Closed
Passed re-test
A section heading skipped a level1.3.1
best practice
ShopSet to level 2
Re-test: No skipped level in the outline
Closed
Passed re-test
A large title repeated the page’s main heading as a second heading—
best practice
Content pageTaken out of the heading outline; still visible
Re-test: Outline shows the main heading only
Closed
Passed re-test
The header menu did not tell assistive technology which page you are on1.3.1Site-wideCurrent-page marker added to the active menu link
Re-test: Checked on three pages
Closed
Passed re-test
Four social links in the header open a new tab without saying so3.2.5
AAA, best practice
Site-wide“(opens in a new tab)” added to each link’s name
Re-test: All four names read correctly
Closed
Passed re-test
Hero slideshow images had no useful descriptions1.1.1HomeDescription for each slide, editable by the store in its theme editor
Re-test: 0 violations on the live site
Closed
Passed re-test
A banner image and a recipe image had no descriptions1.1.1HomeDescriptions set in the store’s file library, so they apply in every theme
Re-test: 0 violations on the live site
Closed
Passed re-test
The skip-to-content link did not lead anywhere useful2.4.1Site-widePointed at the main content
Re-test: Reaches the main content
Closed
Passed re-test
Two product names were read aloud with no space between the words1.3.1Home, ShopSpacing restored in the markup
Re-test: Checked in the browser’s accessibility tree
Closed
Passed re-test
An invalid email in the newsletter form showed an error that screen readers did not announce3.3.1, 4.1.3Site-wide footerError now shown in a status line that is announced
Re-test: Invalid email tested; a valid email was not submitted
Closed
Passed re-test
A tagline was marked up as a heading1.3.1HomeHeading markup removed
Re-test: 0 violations
Closed
Passed re-test
Review badge text measured 4.23:11.4.3ProductDarkened to 5.3:1
Re-test: 0 violations
Closed
Passed re-test
The main page heading was hidden from everyone on four pages1.3.1, 2.4.6Four content pagesHeading made available to assistive technology
Re-test: Heading present on all four
Closed
Passed re-test
The brand heading ran two words together1.3.1HomeSpace restored
Re-test: 0 violations
Closed
Passed re-test
A wholesale note measured 4.48:1 at 14 px, just under the 4.5:1 minimum1.4.3Store locatorSame warm gray one shade darker: 4.72:1
Re-test: An 11-page scan with the full rule set, best-practice rules included: 0 violations on every page
Closed
Passed re-test
The product review widget brings its own colors, and two fell short: review dates and rating counts at 4.23:1, and “Verified” links at 3.39:11.4.3Product pagesBoth darkened within the same hue: 5.3:1 and 5.1:1
Re-test: 0 violations on the 11-page scan. The widget belongs to a third party, so it is re-checked after each of its updates
Closed
Passed re-test
White text on the brand yellow in the About page slider measured 1.68:1. The site’s own navy would only have reached 4.01:1 and still failed1.4.3AboutText changed to a deeper navy from the same family: 6.68:1. Limited to the slider, so no other yellow panel changed (approved by the owner)
Re-test: 0 violations on the 11-page scan
Closed
Passed re-test
The cart icon’s click area measured 45 × 23 px, one pixel under the 24 px minimum2.5.8Site-wide headerFour pixels of padding above and below; the icon did not move
Re-test: 0 violations on the 11-page scan, target-size rule included
Closed
Passed re-test
Slideshow previous and next buttons were 20 × 20 px and 10 px apart2.5.8Site-wideEach enlarged to 32 × 32 px around the same center, so the arrows did not move
Re-test: 0 violations on the 11-page scan, target-size rule included
Closed
Passed re-test
The pause button added for moving content landed on top of a carousel’s Next arrow, so neither could be hit reliably2.5.8Store locatorPause button moved just outside the strip, level with the arrows and 24 px clear
Re-test: 0 violations on the 11-page scan, target-size rule included
Closed
Passed re-test
The “Skip to content” link stayed invisible while focused. The theme’s own stylesheet hid it with rules the platform’s reveal could not override, so a sighted keyboard user tabbed onto something they could not see2.4.1, 2.4.7Site-wideShown on focus as a navy button at top left, white text at 11.3:1, with a yellow focus ring
Re-test: The rule is live, but whether the link is visible while focused can only be confirmed by a person using a keyboard. Queued for the manual pass in Section 6
Closed in code
Manual check pending
Four pages hid their main section entirely, which also removed each page’s main heading from the accessibility tree1.3.1, 2.4.6Four content pagesThe section stays in the accessibility tree and is hidden from view instead, so the heading is announced and nothing changes on screen
Re-test: 0 violations on the 11-page scan; no missing page heading on any of the four
Closed
Passed re-test
The newsletter pop-up and the cart drawer were announced as dialogs but did not behave as dialogs: focus never moved inside them, so Escape never worked2.1.2, 2.4.3, 4.1.2Site-wideEach dialog now moves focus inside when it opens, keeps Tab inside, closes with Escape or a click outside, and returns focus to the button that opened it, whichever code opened it
Re-test: Live, but focus trapping and focus return cannot be confirmed by an automated scan. Queued for the manual pass in Section 6
Closed in code
Manual check pending
Markup repeated across many sections had naming faults no single edit would reach: pasted-in icon files repeated the same ids, icons inside labeled buttons were read twice, pop-ups had no name (the platform’s own privacy banner among them), and embedded frames had no title1.1.1, 4.1.1, 4.1.2Site-wideA pass that runs on every page, including content loaded later such as product panels and the cart, removes duplicate ids, hides decorative icons, names dialogs and titles frames. A later revision cut about 2.5 seconds of processing on a phone
Re-test: 0 violations on the 11-page scan. The code is dated; the exact publish day was not recorded, so this date is the earliest publish the record can show, and is stated as such
Closed
Passed re-test
Controls that responded to a mouse click could not be reached or used by keyboard: quick-add and add-to-cart controls, the product details accordion, and carousel tiles2.1.1, 4.1.2Site-wideEach now behaves as a button: reachable with Tab and working with Space. Accordion headings report whether they are open
Re-test: 0 violations on the 11-page scan. Space tested directly on the live add-to-cart control, with the click suppressed so nothing entered the cart. Same earliest-publish dating as above
Closed
Passed re-test
“Free shipping – Quick add” button text was below the contrast minimum1.4.3Product tilesButton color darkened only as much as needed
Re-test: Live and passing on later scans
Closed
Passed re-test

Section 3

Conformance summary

↑ Contents

When human testing is part of the engagement, a full criterion-by-criterion Accessibility Conformance Report on the VPAT® form is issued after that testing (Section 6), because many WCAG criteria can only be confirmed by a person. This interim summary shows the automated status. It is a report of findings, not a certification.

MeasureResult
WCAG 2.2 criteria in scope55 (31 Level A, 24 Level AA)
Criteria with findings at baseline1.1.1 Non-text content; 1.4.3 Contrast; 2.5.8 Target size; 4.1.2 Name, role, value
Automated violations at latest re-test0 on 11 pages of the live site: home, shop and about pages, two product pages, cart, search, privacy policy, the 404 page and two content pages
Checked by hand on desktopKeyboard order and visible focus on the home page; cart drawer and cookie dialog by keyboard; text over photographs; slideshow controls; linked documents (none found)
Not yet testedReal screen readers (VoiceOver, NVDA, JAWS); phones, including the mobile menu; reflow at 320 px and zoom to 200%

Section 4

Public accessibility statement

↑ Contents

The statement was approved by the store owner and then published, linked from the footer of every page. It claims no conformance. It names WCAG 2.2 Level AA as the target, describes what has actually been done, and lists known limits, including that testing with screen readers is still under way. It will be updated with the results of that testing.

ItemValue
Accessibility contactAccessrank receives reports and replies for the store, with a copy to the store owner
Reply time published2 business days

Accessibility inquiry log

Date receivedChannelIssue raisedResponseResolved
——No inquiries recorded yet——

Section 5

Ongoing monitoring record

↑ Contents

Each scan is recorded with its date and result. Scores from different tools are listed separately and are never added together. A signed monitoring certificate (Form A2) is issued after each scan.

MeasureBefore fixesAfter fixes
Lighthouse accessibility88100
Accessrank score (accessibility)Not run before fixes98, no WCAG failures found
Automated violations4 failed checks, home page0 on 11 pages
On-page SEO (reported separately)—100, 14 of 14 checks
AI agent readiness (reported separately)—100, 8 of 8 checks

Scan log

StageToolResult
BaselineGoogle LighthouseAccessibility 88; 4 failed checks on the home page
After the first roundGoogle LighthouseAccessibility 93
After the second roundGoogle LighthouseAccessibility 100, confirmed on desktop and mobile
After the second roundAccessrank re-test of 57 pages62 speed and stability issues confirmed fixed; contrast and naming items still open and carried into the next round
After the third roundAutomated scan (axe-core)0 violations on the pages scanned
After the fourth roundAccessrank free checkAccessrank score 98; 0 of 55 criteria failing; on-page SEO 100; AI agent readiness 100
After the fourth roundAutomated scan (axe-core), full rule set1 contrast finding on product pages, fixed and published the same day; then 0 violations on 11 pages
After the fourth roundGoogle LighthouseAccessibility 100, performance 92, best practices 100, SEO 100

Section 6

Manual and assistive-technology testing

↑ Contents

In progress. Automated scans cannot confirm every WCAG criterion. A person using a keyboard and a screen reader checks what a scan cannot, such as whether a page makes sense read aloud and whether a purchase can be completed without a mouse. A 14-step test script covers the core shopping path: finding a product, reading its details, choosing options, adding to the cart and using the cart. When testing is done, this record is updated with the results, the tester’s signed attestation and any new findings, and the accessibility statement is updated to match.

Section 7

Program engagement and scope

↑ Contents
ItemDetail
Target standardWCAG 2.2 Level A and AA
Scopestore.example, customer-facing pages and templates. Platform-hosted checkout and account pages excluded
MethodFixes in the theme code and store content; no overlay or widget
MonitoringContinuous automated monitoring, with a dated certificate after each scan
Store responsibilitiesKeep the accessibility contact monitored; tell Accessrank before theme, app or platform changes; add descriptions to new images

Limits outside the theme

AreaOwnerAction
Checkout and payment pagesStore platformDocumented as a known limitation
Subscription widgetApp vendorIssues reported to the vendor
Product review widgetApp vendorColors corrected on the store’s side; re-checked after vendor updates

Language guidelines and disclaimer

↑ Contents
SayDo not say
Working toward full WCAG 2.2 AA conformanceFully accessible
Audited and monitored by AccessrankCertified accessible, or lawsuit-proof
Fixes are made in the site’s codeProtected by an accessibility widget
Conforms to WCAG 2.2 Level AA on [date], per the attached report (only after manual testing is complete and attested)ADA compliant

Disclaimer. This record reports findings against WCAG 2.2 Level A and AA as of the dates stated. It is not legal advice, does not certify compliance with the Americans with Disabilities Act or any other law, and does not guarantee any outcome in any claim or proceeding.

Exhibits and attestation forms

↑ Contents
ExhibitDocument
ABaseline audit reports
BRemediation log as a spreadsheet, and the owner’s approvals of visible changes
CAccessibility Conformance Report (VPAT®), issued when human testing is part of the engagement
DAccessibility statement as published, with a dated screenshot
EMonitoring certificates, in date order
FManual and assistive-technology testing attestations

Form A2: Monitoring certificate

Sitestore.example
Scan date[Date of scan]
Toolaxe-core, WCAG 2.0–2.2 A/AA plus best practice, live site
Pages scanned11
Result0 violations

I certify that the scan identified above was run on the date stated and that the figures above are the results the tool reported. This certificate reports a scan result. It does not certify conformance with WCAG or compliance with any law.

Signature
Printed name
Date signed