Contact

What makes a website fast? Core Web Vitals and performance in practice

A light-filled view of the Grand Canal in Venice with gondolas, sailing boats and palaces. (J. M. W. Turner, Venice, from the Porch of Madonna della Salute, c. 1835)
J. M. W. Turner: Venice, from the Porch of Madonna della Salute, c. 1835 · The Metropolitan Museum of Art · Public domain
Category
Web
Published
Reading time
~4 min read
Author

Summary

A fast website is not the result of one trick but of many small decisions. The three Core Web Vitals show where visitors are losing time.

Why speed is a business issue

Visitors do not notice your technology stack; they notice whether a page responds immediately or makes them wait. A slow page breeds impatience, and an interface that lags when clicked creates doubt. On mobile devices and weaker connections, that friction is felt even more acutely.

Speed is also a search ranking signal: Google factors Core Web Vitals into its page experience evaluation. Fast performance will not rescue weak content, but between two equally relevant pages, the faster one holds the edge.

What do the three metrics measure?

Core Web Vitals metrics and Google's thresholds for a “good” rating
MetricWhat it measuresGood thresholdTypical cause when poor
LCP (Largest Contentful Paint)How quickly the largest visible content element renders2.5 seconds or lessUnoptimised hero images, slow server response, render-blocking scripts
INP (Interaction to Next Paint)How quickly the page responds to clicks, taps and keystrokes200 milliseconds or lessExcessive or heavy JavaScript blocking the main thread
CLS (Cumulative Layout Shift)How much the layout jumps around while loading0.1 or lessImages without explicit dimensions, late-injected banners or embeds

Google evaluates sites using real-user field data, checking whether the vast majority of visits hit the “good” threshold. A single fast test run on a powerful office laptop proves very little.

What slows websites down most often?

  • Oversized images: high-resolution images saved in legacy formats account for the bulk of page weight on many sites.
  • Third-party scripts: analytics trackers, chat widgets and embedded media: each demands network and CPU time outside your direct control.
  • Unused JavaScript: code that the browser has to download and parse even though the current page never uses it.
  • Slow server response: rebuilding pages dynamically on every request instead of serving pre-rendered or cached content.
  • Web fonts: loading too many font weights and styles, delaying text rendering.
  • Bloated themes and plugins: the hidden cost of a convenience plugin is often extra payload paid on every single page view.

What can be done?

Order of operations matters: start where visitors are losing the most time.

  1. Measure with real-user data: Google Search Console and PageSpeed Insights reveal what actual visitors experience.
  2. Optimise images: serve properly sized images in modern formats and lazy-load anything below the fold.
  3. Pre-render where possible: content that does not change on every request should not be rebuilt on the fly each time.
  4. Audit third-party scripts: remove trackers and widgets that deliver no measurable value, and defer the rest.
  5. Reserve space for media: define explicit dimensions for images and embeds so the layout never shifts as assets load.
  6. Monitor continuously: performance is not a one-off checkbox; every new feature or marketing script can degrade it.

Speed is an architectural decision

Retrofitting speed onto a slow site is much harder than building it fast from day one. Technology choices, page architecture and how content is managed set your performance ceiling early on. That is why we treat performance not as end-of-project polish, but as a core design requirement.

As a client, include performance targets in your acceptance criteria, defined in measurable terms and verified on real mobile devices.

Key takeaways

  • Core Web Vitals measure three things: loading speed (LCP), responsiveness (INP) and visual stability (CLS).
  • Google judges performance using real-visitor field data, not a single lab test.
  • The most common bottlenecks are unoptimised images, third-party scripts and excess JavaScript.
  • Start optimisation by measuring real data and tackle the biggest time sinks first.
  • Speed is a design decision; make it a measurable acceptance criterion.

Frequently Asked Questions

Is a green lab score in PageSpeed Insights enough?

The lab score is a useful diagnostic, but it is a simulated test. What truly counts is what real visitors experience, shown in the field data section of PageSpeed Insights and in Search Console once your site has sufficient traffic.

Will making our website faster improve our search rankings?

Speed is one ranking signal among many and will never replace relevant, high-quality content. However, it directly improves user experience and conversion rates, which makes it worthwhile in its own right.

Can an existing website be sped up, or do we need to rebuild it?

Most sites can be improved significantly by optimising images, trimming third-party scripts and configuring proper caching. Whether a full rebuild makes more sense depends on how severely your current platform and theme constrain performance.

Facing a similar decision? We are happy to talk it through.

Let's talk