bumicerts · lighthouse-performance
Bumicerts Lighthouse Performance Audit
Autoresearch improved the Bumicerts root landing route Lighthouse Performance score from 74 to 90, crossing Lighthouse’s green threshold.
Headline result
Green Lighthouse score reached. The final run scored 90 in the Performance category with FCP 1.1s, LCP 3.6s, Speed Index 1.2s, TBT 1ms, and CLS 0. The baseline was 74 with a 41.7s LCP and 12.5MB transfer weight.
Scoring basis
The eval used Lighthouse’s Performance category for a production Next.js build served locally. Per the Chrome Lighthouse scoring reference, the Performance score is a weighted average of metric scores: FCP 10%, Speed Index 10%, LCP 25%, TBT 30%, and CLS 25%.
/Users/satyam/Desktop/Projects/autoresearch.sh
Artifactassets/lighthouse-final.json
Metric comparison
| Metric | Baseline | Final | Change |
|---|---|---|---|
| Performance score | 74 | 90 | +16 |
| First Contentful Paint | 1,209ms | 1,058ms | -12.5% |
| Largest Contentful Paint | 41,714ms | 3,618ms | -91.3% |
| Speed Index | 3,433ms | 1,201ms | -65.0% |
| Total Blocking Time | 41ms | 1ms | -97.6% |
| Cumulative Layout Shift | 0 | 0 | unchanged |
| Total byte weight | 12,522KB | 1,549KB | -87.6% |
What changed
Root route isolated from marketplace app shell
The root landing page was moved out of the marketplace layout path and rendered with only the minimal root provider path, avoiding unnecessary app-wide providers and shell work on the static landing route.
LCP heading rendered immediately
The hero H1 and inline app icon no longer wait for a Framer Motion entrance animation, reducing the LCP delay while preserving the rest of the hero presentation.
Image and icon payload reduced
Large PNG landing photos were converted to appropriately sized JPEG assets. The favicon was reduced from an oversized ICO to a small 32px ICO, lowering diagnostic transfer weight.
Unused Garamond variants trimmed
The global Garamond font configuration now keeps the normal weights needed by the app while removing unused italic and 600-weight preloads.
Evidence
assets/final-metrics.json
Full Lighthouse JSONassets/lighthouse-final.json
Autoresearch logassets/autoresearch.jsonl
Diff statassets/bumicerts-diff-stat.txt
Findings and caveats
Lighthouse variance was visible
Some micro-optimizations produced noisy or flat scores. The report focuses on the final green run and durable changes that align with the scoring reference rather than one-off benchmark tricks.
Rejected ideas
AVIF conversion, removing image priorities, removing the scroll affordance, replacing the client navbar, and dynamic-importing below-the-fold sections either regressed the score or failed to improve the primary metric.
Reproducing
cd /Users/satyam/Desktop/Projects
./autoresearch.shThe script builds bumicerts-monorepo/apps/bumicerts, starts next start on port 3001, runs Lighthouse against /, and emits METRIC lines parsed by autoresearch.