ShipworkSite inspection
Checks
Every error, explainedGuides →
Pricing Learn Guides

What does a controlled Lighthouse run find?

Lighthouse measures one page under controlled conditions. It complements real-user data: where field data says what happened, Lighthouse says which specific audit failed and how much time each fix would save.

Runs the free audit, which includes this check where it applies to your site. Nothing is stored.

What this check inspects

Shipwork asks Google PageSpeed Insights to run Lighthouse for your page, mobile by default, and reads the four category scores: performance, SEO, accessibility and best practices. It also lists the failing audits with the largest estimated time savings.

What a failure means

A category below 50 is an error and below 90 is a warning. The opportunity audits are informational, but they are the most actionable part: each names a change and the milliseconds it would save.

How to fix it

  1. Work the opportunities in the order Lighthouse ranks them. Largest savings first.
  2. Compress and resize images, and serve modern formats.
  3. Remove or defer unused JavaScript and CSS, especially from apps and pixels you no longer use.
  4. Fix the accessibility failures it reports, since those overlap with how crawlers read the page.

A typical failure, worked through

The setupA store carries three abandoned marketing scripts and a large uncompressed hero image. Performance scores 41 on mobile.

What the check reportsThe check reports the failing performance category and lists reduce unused JavaScript and properly size images as the top opportunities, with estimated savings in milliseconds.

The pointLighthouse names the specific fix and its value. Start with the top opportunity rather than rewriting the theme.

Questions

What is the difference between Lighthouse and field data?
Lighthouse is one synthetic run under fixed conditions. Field data is what real users experienced. Both matter: Lighthouse explains, field data decides.
Does a low Lighthouse score hurt ranking?
The score itself is not a ranking factor. The underlying user experience, measured in the field, is. A low score is a signal to investigate, not a direct penalty.
Why does my score change between runs?
Lighthouse conditions vary with network and server load. Treat the score as a range and focus on the specific audits that fail consistently.

Related checks and guides