Aberio doesn't just test your website—it listens to it. During a 90-second test session, Aberio monitors six distinct "signal types" that reveal different categories of accessibility barriers. Together, these signals paint a complete picture of how accessible your website truly is.
The Six Signals: V (Violations), N (System Noise), Gk (Keyboard Gravity), Gp (Pointer Gravity), U (Ungated Navigation), and K (Keyboard Activation). Each signal measures a different dimension of accessibility and user struggle.
Signal V: WCAG Violations
This is the foundation—traditional code-based accessibility testing. Signal V runs 44 comprehensive WCAG 2.1 Level AA rules against the page's HTML, CSS, and ARIA attributes to identify technical violations.
What Signal V Detects:
- Color contrast failures - Text that doesn't meet 4.5:1 ratio (3:1 for large text)
- Missing alt text - Images without descriptive alternatives
- Invalid ARIA - Incorrect role, state, or property usage
- Form labeling issues - Inputs without associated labels
- Heading structure problems - Skipped levels or missing headings
- Keyboard operability - Missing tabindex, unreachable interactive elements
- Semantic structure - Misuse of HTML landmarks and regions
Signal V uses the industry-standard axe-core library by Deque Systems, the same engine powering tools like axe DevTools and used by thousands of organizations worldwide. This ensures Aberio's violation detection is accurate, up-to-date, and legally defensible.
Coverage: Signal V covers approximately 45-50% of WCAG 2.1 Level AA success criteria—specifically the criteria that can be reliably detected through automated code analysis.
Signal N: System Noise
While you're using a page, is it stable? Or does content constantly shift, appear, and disappear unpredictably? Signal N measures DOM instability through statistical z-score analysis.
What Signal N Detects:
- Unexpected DOM mutations - Elements added, removed, or modified without user action
- Layout shifts - Content movement that disrupts reading or navigation
- Auto-updating content - Live regions that change too frequently
- Animation interference - Motion that obscures or confuses interactive elements
- Timing-based changes - Content that appears or disappears based on timers
Signal N calculates a z-score for each interactive component, measuring how much the DOM around that component changes relative to the page average. High z-scores indicate "noisy" components that create instability for screen reader users and keyboard navigators.
Real-World Example:
A form with a "Submit" button that moves every time an error message appears above it. The button's position shifts, the DOM tree restructures, and screen readers lose context. Signal N detects this instability and flags it as a barrier.
Signal Gk: Keyboard Gravity
Keyboard navigation should feel smooth and predictable. When it doesn't, users exhibit recognizable struggle patterns. Signal Gk measures "keyboard gravity"—the tendency for keyboard focus to get stuck, confused, or lost.
What Signal Gk Detects:
- Ping-pong navigation - Tab forward, then immediately Shift+Tab backward (indicates confusion about where focus went)
- Excessive dwell time - Long pauses before keypress (indicates hesitation or searching for visual focus)
- Refocus attempts - Repeated Tab presses with no apparent progress (potential keyboard trap)
- Reverse-direction navigation - Frequent direction changes suggesting difficulty finding targets
These patterns are measurable signals of struggle. An accessible page typically shows linear, confident Tab navigation with minimal backtracking. An inaccessible page produces chaotic, hesitant keyboard patterns.
Signal Gp: Pointer Gravity
Even sighted users rely on predictable mouse behavior. When click targets move, hover states malfunction, or buttons don't respond as expected, Signal Gp detects the resulting struggle patterns.
What Signal Gp Detects:
- Hover instability - Tooltips or menus that disappear when mouse approaches
- Click reversals - Multiple clicks on the same element (first click failed or wasn't acknowledged)
- Retry patterns - Mouse moving to an element, away, then back (target moved or user unsure if it's clickable)
- Pointer drift - Mouse corrections suggesting poor contrast or unclear affordances
Pointer gravity is particularly important for users with motor disabilities, low vision, or cognitive differences who may struggle with small or unpredictable click targets.
Signal U: Ungated Navigation
During a 90-second test, users should be able to complete their task without encountering unexpected navigation changes or feeling compelled to leave the page. Signal U monitors navigation attempts and context changes.
What Signal U Detects:
- User-initiated navigation - Attempts to click browser back/forward or close the tab
- Unexpected redirects - Page changes context without user request
- Auto-playing media with navigation - Videos or audio that spawn new windows
- Modal interference - Dialogs that block all page interaction without clear escape
High Signal U scores can indicate that the page is confusing, frustrating, or creates barriers that make users want to abandon their task. In accessibility testing, "ungated navigation" often correlates with poor user experience.
Signal K: Keyboard Activation
This signal specifically monitors keyboard trap detection, focus loop identification, and activation failures. While Signal Gk measures struggle patterns, Signal K detects outright keyboard failures.
What Signal K Detects:
- Keyboard traps - Focus enters a component but can't exit using standard keys (Tab, Shift+Tab, Escape)
- Focus loops - Tab navigation cycles between the same 2-3 elements indefinitely
- Activation stalls - Enter or Space on a button/link produces no result
- Modal containment failures - Focus escapes from a modal dialog that should trap it
- Unexpected focus jumps - Focus moves to unexpected locations, breaking logical tab order
Critical Signal: Signal K violations are typically severe barriers that completely block keyboard users from accessing content or functionality. Even a single Signal K detection can significantly lower confidence scores.
How Signals Combine: The Confidence Score
After measuring all six signals, Aberio calculates a single confidence score from 0-100%. This score represents how confident we are that users won't encounter accessibility barriers on this page.
The algorithm weighs signals based on severity:
- Signal K (Keyboard Activation failures) has the highest weight—complete blocks are worst
- Signal V (WCAG Violations) is weighted by impact—contrast failures matter more than missing alt text on decorative images
- Signals Gk and Gp (Gravity) indicate struggle, which is serious but not always blocking
- Signal N (System Noise) is weighted by z-score magnitude—extreme instability scores higher
- Signal U (Ungated Navigation) is a contextual modifier—many navigation attempts suggest deeper problems
The result is a nuanced, evidence-based confidence score that reflects both technical compliance and functional usability.
Why Six Signals Matter
No single signal tells the whole story. A page could pass all WCAG checks (Signal V) but still have keyboard traps (Signal K). Another page might have minor violations but exhibit smooth, predictable behavior (low Gk/Gp scores).
By measuring six distinct dimensions of accessibility, Aberio provides the most complete automated picture possible. You're not just getting a list of code errors—you're getting evidence of real user struggle, quantified and ready to act on.