Skip to content
Audits

Website and web application performance audit

A slow site or app: how a performance audit measures real-world speed (LCP, INP, CLS), finds the causes and estimates the gains, with Google’s thresholds and the difference between a PageSpeed score and real-user data.

By Damien Ladurelle · Updated · 6 min read

Contents
  1. The three metrics that matter
  2. The most common causes
  3. Lab data and real-world data: why both matter
  4. Process and pricing
In short

A performance audit measures how fast a site or application really is for its visitors, compares the results with Google’s thresholds (LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1) and traces the problems back to their causes: images, scripts, theme, server, database. It ends with a list of actions ranked by expected gain and effort.

A slow site loses customers quietly: they do not complain, they leave. A performance audit replaces the feeling (“the site is sluggish”) with measurements, finds the causes and says which to fix first. It applies to a brochure site, an online store or a business application.

The three metrics that matter

Google assesses the loading experience with three metrics, the Core Web Vitals, measured on real visits. A page is considered good when at least 75% of visits meet the thresholds (web.dev, accessed November 17, 2026):

Metric What it measures Good Needs improvement Poor
LCP Time to display the largest visible element 2.5 s or less Up to 4 s Over 4 s
INP Responsiveness after a click or input 200 ms or less Up to 500 ms Over 500 ms
CLS Layout shifts during loading 0.1 or less Up to 0.25 Over 0.25

For a business application, I add server response time and the duration of key actions: opening a customer record, searching for an order, saving a form.

The most common causes

Cause Symptom Usual fix
Oversized images High LCP on mobile WebP or AVIF formats, correct dimensions, lazy loading
Scripts loaded everywhere High INP, a page that “freezes” Remove unnecessary scripts, defer the rest
Heavy theme or page builder Lots of CSS and JavaScript Settings, critical CSS, sometimes a theme change
External fonts Text that appears late, layout shifts Fonts served by the site, preloaded
Undersized server Slow response even without traffic Caching, a recent PHP version, suitable hosting
Cluttered database Slow admin pages Cleanup, indexes, reviewed queries
Elements without dimensions High CLS Reserve space for images and banners

Lab data and real-world data: why both matter

The PageSpeed score simulates an average phone on a slow network: it helps reproduce a problem. Real-world data, collected from your visitors’ browsers, tells you what they actually experience. A site can have an average score and good real-world data, or the opposite. The audit relies on both.

For the WordPress sites I look after, I built a performance plugin that measures LCP, INP and CLS on real visits, without cookies, and checks visually that an optimization breaks nothing before keeping it (see NooxPerf). The digital-hipster.com site scores 100 in all four PageSpeed Insights categories, on mobile and desktop.

Process and pricing

  1. Measurements: key pages, mobile and desktop, lab and real-world data.
  2. Diagnosis: the cause of each slowdown, not just the symptom.
  3. Action plan: each fix with its expected gain and effort, so you start with the most profitable ones.

Performance analysis is part of the express audit at $650 (€589)excl. VAT. The full audit at $2,200 (€1,990)excl. VAT combines it with code, security, database and mobile, with a costed plan and a 2-hour debrief. For a WordPress site, see also the WordPress site audit.

Frequently asked questions

My PageSpeed score is poor: is that serious?

Not necessarily. The PageSpeed score is a lab test on a simulated phone. What matters to Google and to your customers is real-user data, shown at the top of the report when your site has enough traffic. The audit looks at both and explains the gaps.

Does speed affect search rankings?

Yes, Core Web Vitals are among the page experience signals Google uses. But they mostly separate pages with comparable content. Their clearest effect is elsewhere: a slow site loses visitors and sales before the page even appears.

Does a caching plugin solve the problem?

It helps a lot on a content site, and hardly at all on an application whose pages are personalized. A badly configured cache can also break the layout or the cart. The audit says what can be cached and what must be fixed at the source.

What gains can I expect?

It depends on the causes. Replacing oversized images or removing unnecessary scripts often brings quick, visible gains. Server or database problems take more work. The audit estimates each action so you can choose.

Not sure about your project?

Let’s talk for 30 minutes: I’ll tell you honestly which option I would choose in your place.

Appearance
Book a 30-minute call Version française