Each preconnect eagerly opens a TCP+TLS connection. On slow networks or low-end devices, too many parallel connections cause head-of-line blocking and starve the connections that matter for LCP. Trim to the 4-6 highest-value origins (the ones serving render-blocking or LCP resources).
Info::
i
14 dns-prefetch redundant with preconnect on same origin
preconnect already does the DNS lookup; adding dns-prefetch to the same origin is at best a no-op. Sample: https://cdn.cookielaw.org, https://assets.adobedtm.com, https://api.company-target.com (+11 more). Remove the redundant dns-prefetch entries.
Info::
i
14 URL(s) appear in multiple hints
The same href shouldn't appear in multiple hint rels -- one wins, the others are wasted markup and may confuse cache behavior. Sample: //target.accenture.com/ (preconnect+dns-prefetch), //play.vidyard.com/ (preconnect+dns-prefetch), //accenture.demdex.net/ (preconnect+dns-prefetch) (+11 more).
Info::
✓
Page uses 30 resource hint(s)
Current Resource Hints
preconnect
+4
14
preload
2
dns-prefetch
+4
14
prefetch—
30 resource hints configured
Each preconnect eagerly opens a TCP+TLS connection. On slow networks or low-end devices, too many parallel connections cause head-of-line blocking and starve the connections that matter for LCP. Trim to the 4-6 highest-value origins (the ones serving render-blocking or LCP resources).
Why this matters
Informational: a labeled value pair from the audit.
preconnect already does the DNS lookup; adding dns-prefetch to the same origin is at best a no-op. Sample: https://cdn.cookielaw.org, https://assets.adobedtm.com, https://api.company-target.com (+11 more). Remove the redundant dns-prefetch entries.
Why this matters
Performance issues directly impact user engagement and conversion rates.
The same href shouldn't appear in multiple hint rels -- one wins, the others are wasted markup and may confuse cache behavior. Sample: //target.accenture.com/ (preconnect+dns-prefetch), //play.vidyard.com/ (preconnect+dns-prefetch), //accenture.demdex.net/ (preconnect+dns-prefetch) (+11 more).
Why this matters
Performance issues directly impact user engagement and conversion rates.
B
HTTP/3 (QUIC)
HTTP/3 not advertised
REVIEW
HTTP/3 not advertised
Info::
i
HTTP/3 (QUIC) is not advertised
HTTP/3 isn't advertised via Alt-Svc and the worker didn't negotiate h3. HTTP/3 reduces handshake latency (1-RTT instead of 2-3 RTTs) and is more resilient on lossy connections. Most modern CDNs (Cloudflare, Fastly, AWS CloudFront, Google Cloud CDN) support HTTP/3 with a single config switch -- consider enabling for mobile-heavy workloads.
B
Compression Algorithm
Main response uses gzip; brotli would be ~20% smaller
REVIEW
Main response uses gzip; brotli would be ~20% smaller
Info::
i
Main HTML response uses gzip (brotli would be ~20% smaller)
All current browsers support brotli. Most CDNs (Cloudflare, Fastly, AWS CloudFront, Vercel, Netlify) can switch from gzip to brotli with a single config flag. Origin servers (nginx, Apache) need a brotli module compiled in -- nginx-brotli or Apache mod_brotli.
C
Main HTML Cache-Control
Action
Main HTML max-age=7200 (~120min) is aggressive
REVIEW
Main HTML max-age=7200 (~120min) is aggressive
Warning::
!
Main HTML cached for 120 minutes -- risks stale auth / SPA state
Long max-age on the main HTML document means users may see outdated auth state, SPA shells, or A/B variants until the cache expires. Consider Cache-Control: no-cache (or max-age=300) for HTML and let your hashed assets (JS/CSS/images) carry the long cache instead.
Server-Timing exposes backend timing breakdowns to browser DevTools (e.g., `db: 45ms; render: 120ms; cache: 2ms`). Useful for diagnosing slow pages without backend log access. Most modern frameworks (Next.js, Cloudflare Workers, Fastly) emit it automatically; absence on a managed platform usually means telemetry headers are stripped at the edge.
B
JavaScript Blocking
2 JS blocking issue(s) detected
REVIEW
2 JS blocking issue(s) detected
Warning::
!
8 render-blocking <script src> tag(s) without async/defer
Each `<script src=...>` without `async`, `defer`, or `type="module"` blocks HTML parsing while the browser fetches and executes it. The block lasts the entire round-trip + execution time -- on slow networks this translates directly into LCP delay. Add `defer` (executes after parse, in source order) for scripts that interact with the DOM, or `async` (executes whenever ready) for analytics / independent scripts. Module scripts (`type="module"`) are deferred by default.
Warning::
!
Total JS execution time is 4.0 s -- over the 3.5s budget
Total JavaScript execution (parse + compile + run) across all scripts exceeds 3.5 seconds. On low-end devices that becomes 7-15+ seconds and shows up directly in TBT and INP. Reduce by: tree-shaking unused dependencies, code-splitting (dynamic `import()`), removing or deferring third-party tracking, and replacing heavy frameworks where they're not needed.
C
Green Hosting
Action
Whether the site is served from green-energy infrastructure
REVIEW
Green Hosting
No green hosting detected
A+
Third-Party Impact
14% third-party, 0 ms blocking
PASS
14% third-party, 0 ms blocking
Info::
i
Third-party code accounts for 14% of page weight (209.0 KiB of 1.4 MiB)
Info::
✓
Third-party blocking time is low (0 ms)
86%
First-party Third-party
A+
Text Compression
All text resources are compressed
PASS
All text resources are compressed
Info::
✓
All text resources are compressed
All text resources are properly compressed.
A
Image Optimization
19 images, 0 KB saveable
PASS
19 images, 0 KB saveable
Info::
✓
All images are well-optimized
19images10 KB
0oversized-0 KB
2legacy format
0missing dimensionsCLS risk
A+
Font Loading
4 fonts (159 KB)
PASS
4 fonts (159 KB)
Info::
i
4 font(s) use font-display: swap (FOUT risk but functional)
Web fonts
4
159 KB total
Render-blocking
0
of 4
Dominant font-display
swap
Most common across fonts
Font loading timeline
TransferFOIT (block)FOUT (swap)
graphik-medium-web.woff2woff2swap
Size40 KB
Load time61 ms
Start127 ms
RiskFOUT — text flashes from fallback to web font
graphik-regular-web.woff2woff2swap
Size49 KB
Load time72 ms
Start128 ms
RiskFOUT — text flashes from fallback to web font
icont.woff2woff2swap
Size17 KB
Load time9 ms
Start310 ms
RiskFOUT — text flashes from fallback to web font
graphik-semibold-web.woff2woff2swap
Size53 KB
Load time31 ms
Start889 ms
RiskFOUT — text flashes from fallback to web font
Optimization checklist
Preload critical fonts (priority=high)
Use woff2 format for all fonts
Set font-display to swap, optional, or fallback
Subset large fonts (≤100 KB each)
A+
Resource Caching
All resources properly cached
PASS
All resources properly cached
Info::
✓
No caching issues found
All static resources have appropriate caching headers.
A+
Critical Rendering Path
No render-blocking resources
PASS
No render-blocking resources
Info::
✓
No render-blocking resources detected
A+
LCP Image Preload
LCP preload audit not available
PASS
LCP preload audit not available
Info::
✓
LCP image preload audit not available for this scan
A+
Server Response Intelligence
2 server-response signal(s) detected
PASS
2 server-response signal(s) detected
Info::
✓
`Vary` header declared: Accept-Encoding
The page declares a `Vary` header, telling downstream caches which request headers the response varies on. Critical for content-negotiated responses (compression, language, cookies, device class).
Info::
✓
Both `ETag` and `Last-Modified` present -- efficient revalidation supported
The page returns both `ETag` (content hash) and `Last-Modified` (timestamp) headers. Browsers can issue conditional GETs with `If-None-Match` or `If-Modified-Since`; the server returns 304 Not Modified for unchanged content, saving bandwidth. ETag is the more reliable of the two (timestamps fail in DST transitions and clock skew).
407 ms
1.6 s
3.3 s