Website Performance Audit: The Step-by-Step Method

Table of Contents

• Before you start: what you need to run an audit
• Step 1, Measure your pages' real speed
• Step 2, Read your Core Web Vitals without the jargon
• Step 3, Test mobile and the conversion path
• Step 4, Check the basic technical setup and SEO
• Step 5, Prioritize fixes with an impact/effort matrix
• Free or paid tools: do you really need to pay to run an audit?
• Frequently asked questions

Before you start: what you need to run an audit

A performance audit doesn't require any paid software or a single line of code. A Chrome browser and access to your site are enough to get started, and everything else runs on free tools you can use from any computer.

I regularly open client files where the first roadblock is mental: the belief that you need a "technical expert" to measure your own site. Over five years of analyzing web accounts, I've seen the opposite. 80% of speed problems show up in a few clicks, without ever opening the source code.

What access and free tools do you need?

Three free tools cover an entire first reliable audit. Google PageSpeed Insights analyzes a page in seconds and combines lab data with real-world data. Lighthouse, built directly into Chrome (right-click, "Inspect," Lighthouse tab), simulates a full page load. GTmetrix rounds out the picture with a breakdown of the heaviest files.

As for access, keep your admin dashboard handy (WordPress, Shopify, PrestaShop) to check your settings, and ideally Google Search Console if you already have it connected. No credit card, no heavy install.

What does a performance audit actually measure?

A performance audit evaluates three things: loading speed, technical stability, and the experience your real visitors perceive. The distinction matters, because a site can post a great lab score while still frustrating your actual customers on a 4G phone.

In practice, you measure how long your page takes to become usable, how responsive it is to clicks, and how much weight is coming from the elements slowing everything down. Together, these signals tell you whether your site is working for you or against you. What follows breaks each measurement down into simple actions.

Step 1, Measure your pages' real speed

PageSpeed Insights shows two separate scores depending on the device: always start with mobile. It's the starting point for any audit, because speed drives both SEO and conversion.

Paste your homepage URL into the tool, run the analysis, then repeat it for your best-selling product page and a category page. Three pages are enough for a first diagnosis. Note the overall score (out of 100) and, more importantly, the detailed timings shown at the top of the report.

What I repeat on every engagement: never stop at the colored score. A 72/100 on mobile can hide a 4-second main content load that's quietly killing your sales. The score is an average; the detailed metrics are what actually speak to your business.

How do you test your site's loading speed?

Run PageSpeed Insights on your key pages and note the loading time of the main content area. Run the test at several points during the day, because shared hosting slows down at peak times.

Round it out with GTmetrix to see the file "waterfall": it shows you exactly which image or script is blocking rendering. To go further on reading these indicators, my full 7-step website audit method breaks down every pillar to measure.

What's the difference between lab speed and real visitor speed?

Lab speed is a simulation. Field speed, on the other hand, measures what your actual visitors experience. This distinction changes the whole audit. Lighthouse tests your page under ideal, controlled conditions, while the Chrome User Experience Report aggregates real page loads from millions of Chrome users.

The classic trap: a Lighthouse score of 90 alongside disastrous field data, because your customers are browsing on 4G with an entry-level phone. In PageSpeed Insights, the top section of the report ("Real-world user experience data") always outweighs the bottom section (the simulated test). Look at the real data first.

Step 2, Read your Core Web Vitals without the jargon

Since 2021, Google has used three specific measurements to judge your site's experience: display speed, click responsiveness, and visual stability. They directly affect both your ranking and your conversion rate.

According to the official web.dev documentation on Core Web Vitals, these signals have been part of the ranking criteria since the "Page Experience" update. A slow site is therefore penalized twice: it converts less and ranks lower in results.

What do LCP, INP, and CLS actually mean?

LCP (Largest Contentful Paint) measures the time before the largest image or main text block appears. That's perceived speed. INP (Interaction to Next Paint) measures the delay between your click and the page's visible reaction, i.e. responsiveness. CLS (Cumulative Layout Shift) measures whether your elements "jump" while the page loads.

An image that shrinks a button at the last second, and your visitor clicks the wrong spot: that's CLS. An "Add to cart" button that takes a full second to respond: that's INP. A homepage visual that crawls: that's LCP. Three symptoms, three different fixes.

What thresholds does Google consider good?

Google sets precise thresholds that give your audit a target. Aim for them page by page.

The first threshold to target is still LCP under 2.5 seconds: it weighs the most on the perception of slowness. If even one metric falls out of the green, Google considers your page "not compliant." To track these indicators over time, the dedicated Search Console report centralizes them, as I detail in this guide to managing your SEO with Search Console.

Step 3, Test mobile and the conversion path

According to Médiamétrie, mobile accounts for more than half of all time spent online in France. Auditing the desktop version first means auditing the wrong front door. In PageSpeed Insights, always switch to the "Mobile" tab before looking at desktop.

Google has indexed your site based on its mobile version ever since the shift to "mobile-first indexing." A site that's fast on desktop but slow on mobile loses sales every day, and those losses never show up anywhere in your regular reporting.

Why audit the mobile version first?

On an e-commerce project with about a dozen employees that I worked with last year, the desktop score looked fine, but mobile LCP reached 4.1 seconds. Visitors were leaving before they even saw the products.

After optimizing images and turning on caching, mobile LCP dropped to 2.3 seconds. Over three months of tracking: conversion rate up by nearly two points, cart abandonment down sharply. Mobile speed isn't a technical footnote; it's direct revenue.

How do you spot where your visitors are dropping off?

Cross-reference your speed data with your conversion reports in Google Analytics 4. Look for high-traffic, low-conversion pages: they're often the slowest ones. A tool like Hotjar or Matomo also lets you see where visitors stop scrolling or click on nothing.

The habit that changes everything: connecting every slow page to its actual conversion rate. To dig deeper into this, my web data analysis method for deciding fast explains how to go from measurement to decision without drowning in numbers.

Redoing this analysis by hand every month takes time few business owners have. Lysible runs it continuously and flags you when an indicator drifts outside its normal range.

Step 4, Check the basic technical setup and SEO

Four technical bottlenecks explain most slow sites: oversized images, no caching, too many third-party scripts, and chained redirects. None of them require coding to spot.

Across the accounts I audit, uncompressed images come out on top, by far. A 2 MB image on the homepage is the most common mistake, and the easiest to fix. PageSpeed Insights lists exactly which files to shrink in its "Opportunities" section.

Here's the technical checklist built for a small business, no code required:

• Image weight: compress any image over 200 KB, switch to WebP, and resize to the actual display dimensions.
• Caching: turn on browser caching via a plugin (on WordPress, WP Rocket or similar) to avoid reloading the same files on every visit.
• Third-party scripts: every widget (chat, customer reviews, ad pixels) slows the page down. Remove the ones you no longer use.

These three actions cover most of the quick wins. The fourth, often overlooked: redirect chains. Each redirect hop costs about half a second, and it's not unusual to find three redirects stacked on internal links left over from an old redesign. Check your key links; it's ten minutes of work for immediate gains.

What technical problems slow down performance?

The most common culprits are heavy images, ad scripts, and undersized hosting. GTmetrix shows the total page weight and lists the biggest files, so you can spot them without a developer. A mobile page weighing more than 3 MB is a clear warning sign.

On the basic SEO side, check that your key pages are properly indexed in Search Console and that your sitemap file is up to date. Technical performance and visibility move together: a complete audit connects the two, as in my automated SEO audit method for small businesses.

Step 5, Prioritize fixes with an impact/effort matrix

After an audit, the trap is paralysis: twenty problems listed, no order, and ultimately nothing gets fixed. Ranking each fix by its business impact and the technical effort it requires is the step everyone skips, and it's exactly the one that turns a report into results.

The solution fits into a four-quadrant matrix. Simple, quick to fill in, and remarkably effective for deciding what to do first thing Monday morning.

What should you start with when everything feels urgent?

Always start with high-impact, low-effort actions: those are your quick wins. Sort each fix into one of the four quadrants below.

Compress your images and turn on caching first: these two quick wins often shave a full second off LCP in half a day. Save the big projects (switching hosts, a redesign) for a planned moment. Skip, without guilt, anything touching pages with no traffic or value. This discipline ties every action back to a concrete result on your sales, in line with the marketing KPIs to track when you run a small business.

Free or paid tools: do you really need to pay to run an audit?

For a small business, free tools cover about 90% of the need. Paying only makes sense for continuous, automated monitoring. It's a position I hold against the grain of agencies that push expensive suites from the very first audit.

PageSpeed Insights, Lighthouse, GTmetrix's free version, Search Console, and Matomo form a complete arsenal for measuring, understanding, and fixing. I've verified across dozens of accounts that an owner who knows how to combine these tools spots as many problems as a consultant billing by the day.

Are free tools enough for a small business?

Free tools are more than enough for a one-off audit, as long as you know how to combine them. The limit shows up elsewhere: ongoing tracking. Testing three pages once is free; monitoring a hundred pages every week and getting an alert when LCP slips quickly becomes unmanageable by hand.

Start free. Only invest in a paid tool the day recurring monitoring costs you more time than it saves you. For a small business, that day comes later than people tell you.

Turning your performance audit into an ongoing system with Lysible

A one-off audit gives you a snapshot, but a site's performance shifts every week: an update, a new script, a heavy image, and your metrics slip. The real challenge for a small business is moving from an isolated test to clear, continuous monitoring, without spending your evenings on it. That's exactly what Lysible is built for: automatically cross-referencing your speed, Core Web Vitals, conversion, and SEO data in a single clear dashboard. You see at a glance what's slowing down your sales, without juggling five tools or depending on an agency. The audit becomes a habit, not an annual chore.