Free Tool

PageSpeed Analyzer

Find out exactly how fast your website loads, and what's slowing it down. We target 90–100 on every site we build. The industry average is 30–65.

Tests mobile + desktop simultaneously  ·  Same signal Google uses for rankings

Compare with a competitor (optional)

Their scores will appear alongside yours in the results.

Why do different tools give different scores?

These tools aren't inconsistent; they're each measuring something different. None of them are wrong. They just answer different questions.

Composite Score

HubSpot Website Grader & similar

Combines performance, SEO, mobile, and security into a single weighted score using their own formula. Useful overview, but the number is proprietary, not a Google ranking signal.

Answers: What's my overall web presence health?

Link Authority

Moz / Semrush / Ahrefs

Indexes your backlinks and calculates domain authority based on who links to you. Measures off-page reputation and competitive positioning, completely separate from site speed or technical health.

Answers: How authoritative is my domain on the web?

HTML Audit

SEO checkers & site auditors

Crawl your HTML and check for meta tags, heading structure, broken links, and on-page elements. They read your source code, they don't run your page, so they can't accurately measure load speed.

Answers: Is my page structured correctly for search engines?

Load Time

GTmetrix / Pingdom

Measure raw load time from their own fixed server locations. Useful for waterfall analysis, but different test conditions than Google (different device, different network, different city), so the numbers naturally differ from PSI.

Answers: How long does my page take to fully load?

Why does the same tool give a different score each time?

Getting a 91 on Monday and a 78 on Friday doesn't mean something broke. Five factors cause this, and all of them are normal.

Google runs a live test every time

There's no cached score. Every run triggers a real Lighthouse audit, so conditions at that exact moment affect the result.

Network simulation is probabilistic

The throttled 4G connection is simulated, not a fixed cable. Slight fluctuations in the simulation change how assets load and when they're measured.

Third-party scripts load unpredictably

Analytics, chat, and booking widgets call external servers on every test. If those servers are slow that day, or blocking, your score drops, even though your site didn't change.

CDN and server cache state varies

A cold cache (first visitor after a quiet period) loads slower than a warm one. If the test hits your server before content is cached, load time goes up.

Google's own infrastructure fluctuates

The servers Google uses to run tests aren't always under identical load. A busy period on their end can add latency that shows up in your score.

What this means in practice

Google officially acknowledges ±5–15 point variance as normal. A score of 85–95 across multiple runs means the same thing: a fast site. Trust the range, not any single number. If you see a consistent drop over weeks, then something changed.

What do these scores mean?

90–100

Fast

Google ranks fast sites higher. Visitors stay, convert, and come back.

50–89

Needs Work

Visible lag on mobile. You're likely losing visitors before the page loads.

0–49

Slow

Actively hurting your rankings and conversions. Google penalises slow sites.