Context

The quality.arc42.org knowledge base serves software architects and developers who browse ~220 quality attributes, ~140 requirements, and an interactive force-directed graph. Users typically arrive via search engines and decide within seconds whether the site feels trustworthy and worth exploring. Visual appeal directly affects engagement and return visits.

Trigger

A new or returning user loads any page of quality.arc42.org on a desktop or mobile browser.

Acceptance Criteria

  • At least 70% of surveyed users (n >= 50) rate overall visual appeal >= 4 on a 5-point Likert scale, measured via the VisAWI questionnaire (Visual Aesthetics of Websites Inventory) administered annually. Assumption: 70% threshold chosen conservatively for a non-commercial knowledge site.

  • AttrakDiff hedonic quality score (HQ-I subscale) reaches at least 0.5 on the -3 to +3 scale across a panel of >= 30 representative users, evaluated per major redesign via online self-report.

  • Visual consistency: no more than 5% of sampled pages (random sample of >= 30 pages per audit) deviate from the site’s color scheme, typography, or spacing conventions, verified by automated visual regression testing per release.

  • Cumulative Layout Shift (CLS) remains <= 0.1 on at least 95% of page loads over a rolling 30-day window, measured via real-user monitoring or PageSpeed Insights.

  • System Usability Scale (SUS) overall score >= 68 (industry average) across >= 20 users, evaluated annually.

Monitoring Artifact

Annual UX survey dashboard combining VisAWI scores, AttrakDiff profiles, SUS scores, CLS telemetry from PageSpeed Insights, and bounce rate trends from web analytics.

Measurement instruments referenced above:

  • Likert scale — a symmetric agree/disagree rating scale (typically 5 or 7 points), widely used in survey research to quantify subjective attitudes.
  • VisAWI (Moshagen & Thielsch, 2010) — a validated 18-item questionnaire measuring perceived visual aesthetics of websites across four facets: simplicity, diversity, colorfulness, and craftsmanship.
  • AttrakDiff (Hassenzahl et al., 2003) — a semantic differential questionnaire measuring pragmatic quality (usability), hedonic quality (stimulation and identity), and overall attractiveness of interactive products.
  • SUS (Brooke, 1996) — a 10-item questionnaire producing a single score (0–100) for perceived usability. Scores above 68 are considered above average.