A faster website should help people read, navigate and complete tasks with less delay. Diagnose the bottleneck on representative pages before choosing a fix. A performance score is a useful signal under stated conditions, not a complete measure of customer experience or a promise of better search rankings.

This guide covers page bottlenecks, media and script costs, representative tests and performance evidence. Discuss the relevant requirements through ITZ’s Website Speed Optimisation service.

Loading and interaction metrics

Largest Contentful Paint measures when the largest visible content element renders. Interaction to Next Paint measures interaction responsiveness. Cumulative Layout Shift measures unexpected visual movement. Interpret them together rather than treating a single score as the whole user experience.

Use the metrics to frame specific problems. LCP concerns loading of the largest content element, INP concerns interaction responsiveness and CLS concerns unexpected movement. Current Web Vitals guidance identifies good thresholds of 2.5 seconds, 200 milliseconds and 0.1 respectively, evaluated at the 75th percentile. These are diagnostic targets, not a guarantee that every page or visitor has a good experience.

Field and laboratory evidence

Field data reflects eligible real-user visits, while laboratory tests use selected conditions. Differences can arise from device capability, connectivity and interaction patterns. Record test conditions and compare equivalent pages and journeys.

Distinguish real-user observations from controlled tests. A laboratory run can help identify expensive scripts or a slow hero asset, while field data reflects the eligible visits available to the reporting system. Low traffic may leave insufficient field evidence. Record device, connection and page conditions before comparing results, and avoid interpreting one favourable run as a lasting improvement.

Request a Website Performance Review to discuss page bottlenecks, media and script costs, representative tests and performance evidence.

Technical improvements

Optimise large images and fonts, reduce unnecessary scripts and review rendering work. Reserve space for media and dynamic elements. Prioritise changes according to the diagnosed bottleneck and required site functionality.

Investigate image delivery, font loading, JavaScript execution and server response according to the evidence. Set dimensions for media and review third-party tools that consume resources or move content after load. Removing a business-critical function without discussing the trade-off is not an acceptable optimisation. Test the affected journey after each meaningful change.

Validation and search

Retest after changes and monitor field data when sufficient traffic is available. Good page experience supports users, but performance alone does not determine search rankings or guarantee visibility.

Google describes Core Web Vitals as part of page experience, with good scores not guaranteeing high search placement. Report what changed technically and what measurements show. If a goal concerns enquiries or sales, those outcomes need separately verified data. Keep the scope clear when one page template improves but other site sections have different bottlenecks.

A practical example

Illustrative example: a homepage has a large image as its main loading element, while a product page is delayed by third-party scripts. Applying the same image compression everywhere may help one problem and miss the other. Compare page types and prioritise the bottleneck supported by the diagnostic evidence.

Your review checklist

  • Select representative pages and record test conditions.
  • Use field and laboratory evidence for their appropriate purposes.
  • Prioritise diagnosed bottlenecks while preserving required functions.
  • Retest journeys and report technical results separately from sales.

Frequently asked questions

Why can a laboratory score differ from field data?

The two measure different conditions and populations. Use laboratory tests to diagnose issues and field data to understand the experience observed across actual visits.

References and further reading

Related services and guides

Discuss the next step

Request a Website Performance Review. Use the form below to discuss page bottlenecks, media and script costs, representative tests and performance evidence. Share the relevant goals and existing arrangements; do not include passwords, private records or access tokens.

Request a Website Performance Review ↗

Reviewed and expanded on 15 October 2026. Examples are illustrative; the appropriate approach depends on your business requirements and existing systems.