WebsitesModule 4: Quality and inclusionLesson 12 of 15
Course progress73%

16 min lesson · Updated August 2026

Why does website speed matter?

Website performance is how quickly and smoothly people can receive, see and interact with useful content; it depends on servers, networks, assets, code, third parties, devices and page design—not one score.

What you will learn

By the end, you will understand:

  • Explain loading, responsiveness and stability
  • Understand LCP, INP and CLS in context
  • Use field and lab evidence to diagnose improvements

Visual explainer

See the idea clearly.

Performance is experienced over time

  1. 01

    Request begins

  2. 02

    Server responds

  3. 03

    First useful content

  4. 04

    Main content appears

  5. 05

    Page becomes responsive

  6. 06

    User interacts

  7. 07

    Feedback appears

  8. 08

    Later content stays stable

Core Web Vitals

MetricMeaning
LCPHow quickly the largest relevant viewport content element appears.
INPInteraction responsiveness across the page visit, measured from input to next paint.
CLSUnexpected visual layout movement during the page lifespan.

Field and lab evidence differ

Field

Aggregated eligible real-user data over time; reflects diverse devices/networks but may be unavailable or broad.

Lab

Controlled synthetic test for diagnosis; repeatable but not every real experience.

Common cost centers

  • Slow server/redirects
  • Oversized images/video
  • Render-blocking CSS/fonts
  • Large/long JavaScript
  • Third-party scripts
  • Unstable dimensions
  • Uncached static assets
  • Too many requests
  • Heavy widgets
  • Poor mobile delivery

Optimize meaningfully

Prioritize representative high-value pages and real-user evidence. Compress and size images, reduce unused work, reserve space, cache safely and delay nonessential third parties.

Do not remove accessibility, security or critical function merely to raise a lab score. Test outcomes after changes.

Performance needs a budget and monitoring

Set page-type budgets for bytes, requests, scripts or experience metrics and test them in delivery workflows. Monitor field distributions and regressions.

One fast launch does not prevent later content, tag or plugin growth. Assign ownership.

Real-world example

Example: a hero image is visually small but technically huge

Example

A mobile page displays a 400-pixel-wide crop but downloads a multi-megabyte desktop original. Responsive image sources, modern format, dimensions and appropriate preload improve LCP without removing the visual.

Try this

Find the main delay

Run PageSpeed Insights on a representative URL. Separate field and lab sections, identify the LCP element and one long main-thread task, then verify the actual resource in browser network/performance tools.

Common questions

Questions beginners ask.

Why does website speed matter?

Delays and instability can block tasks, accessibility and trust, especially on constrained devices/networks.

What is LCP?

A Core Web Vital measuring when the largest relevant content element in the viewport renders.

What is INP?

A Core Web Vital measuring overall interaction responsiveness from input to next paint.

What is CLS?

A measure of unexpected layout shifts that affect visual stability.

Is a score of 100 required?

No. Use metrics and diagnostics to improve real experiences, not chase a number detached from purpose.

Why do speed tests vary?

Network, device, server, cache, page state, third parties and tool conditions vary.

Are animations always slow?

No. Efficient transform/opacity animation can be smooth, while layout-heavy or script-heavy effects can be costly.

Should every script be removed?

No. Keep justified functionality and optimize/load it according to priority.

Assessment

Check what you understood.

5 questions · instant explanations

1. What does INP measure?
2. What does CLS measure?
3. Why use both field and lab data?
4. Which often improves image performance?
5. True or false: the highest lab score is always more important than useful accessible functionality.

Sources

Primary references.