Skip to main content
Nerve Centre KPIs · Audit Profile · Sentiment Settings Prismic is a slice-based headless CMS, so what matters to a merchant is the RENDERED front-end, not the content API in isolation. This audit answers: (1) is auth / repository / live-URL access healthy, (2) is the rendered site passing Core Web Vitals (LCP, INP, CLS, TTFB) and holding its Lighthouse Performance / SEO / Accessibility / Best Practices scores, (3) is page weight / render-blocking / blocking-time within budget, and (4) are high-traffic commerce pages rendered from Prismic slices failing CWV and bleeding conversion (the killer cross-channel area).

​
What this audit checks

​
Authentication & access

  • Repository reachable and master ref resolvable (auth on /api/v2)
  • Access token valid for private repositories (else document-level cards degrade)
  • Live site URL reachable and scoreable by PageSpeed Insights
  • PageSpeed Insights API key present and under daily quota

​
Core Web Vitals (field)

  • Origin CWV pass rate < 50% (rendered site failing the field test)
  • LCP p75 > 4000ms (poor) on mobile
  • INP p75 > 500ms (poor) on mobile
  • CLS p75 > 0.25 (poor) - slice / template layout shift
  • TTFB p75 > 1800ms (poor) - slow origin / SSR / edge response

​
Lighthouse lab health

  • Performance score < 50 on mobile (regression after a slice / template change)
  • SEO score < 70 (missing meta / structured-data hygiene on rendered pages)
  • Accessibility score < 70
  • Best Practices score < 70
  • Mobile vs desktop performance gap widening

​
Page weight & opportunities

  • Total page weight > 3MB (un-optimised assets / oversized slices)
  • Total Blocking Time > 600ms (heavy JS hydration)
  • Speed Index > 5800ms (poor)
  • Render-blocking resources with > 500ms savings available

​
Cross-channel: content-to-commerce (the killer area)

  • High-traffic commerce landing/product pages (sibling shopify/bigcommerce/adobe_commerce) rendered from Prismic content failing Core Web Vitals - conversion-impacting
  • Slow Prismic-rendered pages driving sessions (sibling google_analytics) but converting below site average (CWV / merchandising gap)
  • Slice / template change date correlates with a CWV regression on a revenue page
  • Origin CWV pass rate down vs prior period while commerce traffic is flat or up (rising bounce risk)

​
Severity thresholds

SignalWarnCritical
cwv_pass_rate7550
lcp_p75_ms25004000
inp_p75_ms200500
cls_p750.10.25
ttfb_p75_ms8001800
lighthouse_performance9050
lighthouse_seo9070
total_byte_weight15728643145728
total_blocking_time_ms200600

​
Data sources

  • GET https://{repository}.cdn.prismic.io/api/v2 - Repository / master-ref auth + content-type inventory
  • GET https://{repository}.cdn.prismic.io/api/v2/documents/search - Published documents + their live URLs for the content→URL join
  • GET https://pagespeedonline.googleapis.com/pagespeedonline/v5/runPagespeed - Lighthouse lab metrics + scores and CrUX field data for the rendered site