There is no single cliff, but there are three useful bands. Under 2.5 seconds on mobile you are fine. Between 2.5 and 4 seconds you are losing some visitors. Above 4 seconds you are losing meaningful numbers. Most trades sites we test sit well above 4, and the cause is almost always photographs.

One caution up front: site speed is the most oversold item in web design. Moving a site from 2.0 seconds to 1.5 will not change your business, and anyone charging you a retainer to chase that is selling you decimal places. The gains are real only at the slow end, and this article is mostly about telling those two situations apart.

What the famous statistics actually say

Almost every article on this subject cites the same research, and most of them misread it.

The source is a Google and SOASTA study from 2017, which trained a machine learning model on bounce and conversion data from mobile sites. Its headline finding was that as page load time goes from one second to ten seconds, the probability of a mobile visitor bouncing increases by 123 percent. Intermediate figures put the increase at 32 percent going from one to three seconds, and 90 percent going from one to five.

Two things worth understanding about those numbers.

They are relative, not absolute

"Bounce probability increases 32 percent" does not mean 32 percent of people leave. It means whatever your bounce rate was, it goes up by roughly a third in relative terms. If 20 percent of visitors were leaving, about 26 percent now do. That is a real loss and worth fixing. It is not the catastrophe the number implies when quoted carelessly.

Most of the data is ecommerce, not trades

The underlying research largely measured retail and mobile landing pages. Someone idly browsing shoes and someone whose water heater has just failed are not the same visitor, and applying shopping-cart abandonment data to an emergency plumbing call is sloppier than most articles admit.

Why trades sites are genuinely different

The differences cut both ways, which is why a single threshold does not exist.

Your visitor is more motivated. A person with water on the floor will tolerate more friction than a browser comparing running shoes. They need this problem solved and they need it today. That patience buys you a little room.

But your competitors are one tap away. They did not arrive at your site through a considered process. They tapped the third result in a list, and the back button returns them instantly to four alternatives who do the same job. In ecommerce a slow page loses a sale; in local services a slow page hands a customer to a named competitor.

And the connection is often bad. Basements, parkades, older buildings, rural properties, the side of a highway. Your site does not get opened in ideal conditions. It gets opened in the worst ones.

The net effect is that motivation and competition roughly cancel out, and connection quality makes things worse. Which lands you back at the standard thresholds, but with less margin than a typical business has.

The thresholds that actually matter

Rather than raw load time, Google measures three things it calls Core Web Vitals. These are what appear in Search Console and what feed the ranking signal.

MetricGoodPoorWhat it means
LCP≤ 2.5s> 4.0sHow long until the main content appears
INP≤ 200ms> 500msHow quickly the page responds when tapped
CLS≤ 0.1> 0.25How much the layout jumps while loading

For a trades business, LCP is the one that matters most. It is closest to the plain question of how long someone stares at a blank screen before seeing your phone number.

Two details that most explanations skip, and both matter:

Worth knowing that CLS, the layout-jumping one, matters more than it sounds. If your phone number shifts position half a second after the page appears, someone aiming for it taps something else. That is a lost call caused by a metric most owners have never heard of.

Where the real problem usually is

Here is the part that determines whether any of this is worth your attention.

The interesting gap is almost never 1.8 seconds versus 1.5. It is 12 seconds versus 3. Sites in the genuinely bad range are not slightly unoptimised; they are carrying something heavy. In rough order of frequency:

Photographs, by a wide margin

A modern phone produces images of several megabytes. Upload a dozen job photos straight to a gallery page and the page has to move perhaps thirty megabytes before it finishes. On a good connection that is slow. In a parkade it is hopeless.

Properly resized and compressed, those same photos look identical on screen at a fraction of the weight. This single fix resolves the majority of genuinely slow trades sites.

Homepage sliders

A rotating hero loads several large images before the visitor sees anything, to display content most of them never scroll through. It is the worst weight-to-value ratio on a typical site.

Third-party scripts

Chat widgets, review carousels, embedded social feeds, booking tools, and several years of accumulated tracking tags. Each adds a request to a server you do not control, and your page waits for all of them.

Cheap or overloaded hosting

If the server takes a second and a half to begin responding, everything downstream inherits that delay and no amount of image optimisation recovers it.

What to do, in order

The short version

Under 2.5 seconds you are fine and should stop thinking about it. Over 4 seconds you are losing calls and should fix it this month. In between is a judgement call that rarely justifies a retainer.

And if your site is slow, it is probably the photos. Before commissioning anything, have someone check that first, because it is often an afternoon of work rather than a project.

Related: what a plumber's website actually needs covers what belongs on the page once it loads quickly enough for anyone to read it.

Frequently asked questions

How fast should a contractor's website load?

Aim for a Largest Contentful Paint under 2.5 seconds on mobile, measured on real visitor data rather than a desktop test. Between 2.5 and 4 seconds you are likely losing some visitors. Above 4 seconds Google classifies the experience as poor and the losses become significant.

What is Largest Contentful Paint?

Largest Contentful Paint, or LCP, measures how long it takes for the biggest visible element on the screen to appear, usually the main heading or hero image. Google treats under 2.5 seconds as good and over 4 seconds as poor, assessed at the 75th percentile of real visits.

Does site speed affect Google rankings?

Yes, but modestly. Core Web Vitals are a confirmed ranking signal that generally acts as a tie-breaker between pages of similar relevance rather than overriding content quality. For most trades businesses the conversion effect matters more than the ranking effect.

What usually makes a trades website slow?

Photographs, in most cases. Job photos uploaded straight from a phone can be several megabytes each, and a gallery of them will dominate load time. After images, the usual causes are homepage sliders, chat widgets, embedded social feeds, and accumulated tracking scripts.

Is going from 2 seconds to 1.5 seconds worth paying for?

Rarely, for a local service business. The meaningful gains are in moving a genuinely slow site into the acceptable range. Once you are comfortably under the good threshold, further optimisation produces diminishing returns and the effort is usually better spent on reviews or content.

How should I test my own site speed?

Test on a real phone using mobile data rather than office wifi, then check field data in Google Search Console's Core Web Vitals report, which reflects what actual visitors experienced. A fast score on a desktop test tool means little if real visitors are on mid-range phones.

West Coast Media Solutions Inc. builds websites for trades and service businesses across Canada. Speed figures cited are from published Google research and current Core Web Vitals thresholds; your own field data in Search Console is the measure that matters for your site.

Not sure whether your site is actually slow?

We will look at your field data and tell you honestly whether it is worth fixing, including when the answer is that it is not.

Start a Conversation