Speed

Website speed test

Find out how long your page really takes to become usable, what is making it slow, and which fix is worth doing first. Google's own measurements plus our scan of what your page actually weighs.

Free
No account
Usually under a minute

What this checks

Every one of these is measured directly on your page and reported with the evidence attached.

Server response time: How long your server takes to say anything at all. If this is slow, nothing else on the page can start.
Full page delivery: How long until everything has arrived and the page has stopped moving.
Total page size: What a visitor has to download to see your page — the number that decides what happens on mobile data.
Oversized files: The specific files that are too big, named individually, so you know exactly what to compress.
Image format: Whether your images use modern formats. The same picture in WebP is routinely a third of the size with no visible difference.
Compression: Whether your server compresses text before sending it. Switched off, this alone can quadruple what every visitor downloads.
Repeat visit speed: Whether returning visitors download everything again. Invisible to you, because your own browser has it all cached.
Add-ons and scripts: How much code runs on the page, and how much of it blocks the page from appearing.
Outside services: How many other companies' servers your page waits on — chat widgets, fonts, trackers, embeds.
Connection method: Whether your server uses a modern protocol, which changes how efficiently many files can be fetched at once.
Address redirects: Whether visitors are bounced through extra hops before reaching the page. Each one is a round trip they wait through.

These run alongside the rest of the audit — every scan covers all six areas, so you get speed, works on phones, found on google, secure, looks legitimate and easy to contact whichever page you started from.

Two kinds of speed test, and why you want both

Almost every speed tool gives you one of two things, and the two frequently disagree.

Lab measurement loads your page once, now, on a simulated mid-range phone over a throttled connection. It is repeatable, which makes it the right tool for diagnosis: change one thing, run it again, see whether it helped.

Field measurement is what real visitors actually experienced over the past four weeks, collected from real browsers. It is messy, it lags a month behind your changes, and it is the one Google uses when speed feeds into ranking.

A page can score well in the lab and badly in the field — usually because real visitors are on worse devices than the simulation assumes. When the two disagree, the field data is the truth and the lab data is the explanation. This audit reports both where both are available, and says which is which rather than blending them into one reassuring number.

A score is not a diagnosis

The number most speed tools lead with is a weighted composite of several measurements. It is useful for tracking your own progress and close to useless for deciding what to do, because two sites with the same score can have entirely different problems.

What actually helps is the list underneath: which files are too big, named; how much of your page weight is images versus scripts; whether compression is off; whether returning visitors are re-downloading everything. Those are actions. A score is a mood.

So the report leads with the findings and treats the number as a summary of them, not the other way round.

The three findings that come up most

Images that were never resized. A photo straight from a phone is several megabytes and four thousand pixels wide, displayed in a box six hundred pixels wide. The browser downloads all of it and discards most of it. This is the most common finding on this site by a wide margin, and it is the one an owner can usually fix without a developer.

Compression switched off. Text files compress to roughly a quarter of their size, automatically, at almost no cost. It is one server setting, and a surprising number of sites have never had it enabled.

No caching for returning visitors. Your logo has not changed in three years, and without a cache instruction every visitor downloads it again on every page. This is invisible in your own testing, because your browser already has a copy.

None of these three requires a rebuild. Together they account for the majority of the improvement available on a typical small business site.

What speed is actually worth

Two honest claims and one we will not make.

Speed is a confirmed Google ranking signal, and it is a mild one. It breaks ties between comparable pages. It will not lift a page above one that answers the question better, and anyone selling a speed fix as an SEO strategy is overselling it.

Speed's real value is that people leave slow pages. That effect is large, it is measurable in your own analytics, and it does not depend on any search engine's algorithm — which makes it the more reliable reason to care.

What we will not claim is a specific percentage. The published figures on abandonment vary enormously by industry and traffic source, and quoting a precise number from somebody else's e-commerce study as though it applies to your business would be inventing a fact.

Common questions

What is a good website speed?

The main content should appear within 2.5 seconds. Between 2.5 and 4 seconds needs work, and over 4 seconds is costing you visitors before the page has finished drawing.

Why does my speed score change every time I test?

Lab tests run on shared infrastructure over real networks, so consecutive runs vary by a few points. Look at the trend over several runs rather than treating one number as exact, and compare like with like — the same tool, the same page, the same connection type.

My site is fast for me. Why does the test say it is slow?

Two reasons, both common. Your browser has cached most of the site, so you are not downloading what a first-time visitor downloads. And the test deliberately simulates a mid-range phone on a throttled connection, which is closer to your average visitor than your office desktop is.

Will a faster host fix a slow website?

Only if the problem is server response time. If the pages are slow because they carry huge images and a dozen third-party scripts, a faster host delivers those same files slightly sooner. The report tells you which half your problem is in before you pay to move.

Does this test mobile or desktop?

Both are measured where Google has data for them, and reported separately. Mobile is the one to prioritise: it is the majority of traffic on most sites, and it is the version Google indexes and ranks.