Say your site takes seven seconds to load on a phone and a competitor's takes two. You both do good work, and the customer calls whoever's page showed up first.
I spent five-plus years as a developer inside the PPC and SEO industry before starting MAKR Labs. In that time I watched small businesses pay for clicks that landed on pages too slow to convert. The ads did their job and the page lost the customer, and nobody ever billed for that part.
Slow small-business sites are rarely slow for interesting reasons. It's almost always the same four problems stacked on top of each other.
Where the weight comes from#
Page builders. Elementor, Divi, WPBakery and similar builders make editing easy by loading everything on every page. That means slider libraries you never use, an icon font for three icons, and scripts that load whether the page needs them or not. Whoever edits the site gets the convenience, and every visitor pays for it in download time.
Plugin stacks. An aging WordPress site commonly runs twenty to forty plugins, and each one can inject its own CSS and JavaScript into every page. So the gallery plugin loads on your contact page and the form plugin loads on your gallery. Plugins get added when there's a problem and tend to stay installed long after the problem has gone away.
Unoptimized images. A photo straight off a phone is often 3 to 5 MB. Put twenty of them on a job-photos page and you're asking a visitor to download a small movie over parking-lot LTE. Most of those photos could be a tenth of the size with no visible difference.
Cheap hosting. On five-dollar shared hosting, your site waits in line behind hundreds of other people's sites. The server can take a second or more just to start answering, and every other problem on this list adds to that delay.
| Culprit | What it adds | The fix |
|---|---|---|
| Page builder | Scripts and styles for features you never use | A lean build without the builder |
| Plugin stack | Every plugin injects code into every page | Audit them and keep only the few you need |
| Raw images | Multi-megabyte photos on mobile connections | Resize, compress, serve modern formats |
| Shared hosting | A long wait before the page even starts | Hosting that answers in milliseconds |
What a slow site costs#
You don't need a study to know people hate waiting, though the studies exist. Google's own published research, cited everywhere for years, found that the chance of a bounce climbs sharply as mobile load time stretches past a few seconds. I'd treat the exact percentages as rules of thumb. The trend itself isn't in doubt.
For a contractor, it works like this. Most local searches happen on phones. Someone with a burst pipe opens two or three tabs and calls whoever loads first and looks legitimate, so the slow tab loses by default. Your work can be better and it won't matter if nobody waits long enough to see it.
If you run ads, the lost visit is one you already paid for. Google also factors landing page experience into ad quality, so a slow page can quietly raise what you pay per click. You end up paying more for each click while fewer of them turn into calls.
What to ask your developer#
You don't need to become technical. Ask these and pay attention to how clearly they're answered.
- What does the site score on PageSpeed Insights on mobile? Make sure the number you get is the mobile one.
- What's the largest image on the homepage, in kilobytes?
- How many plugins are installed, and what do the bottom five do?
- What would happen to load time if the page builder were removed?
- Where is the site hosted, and how fast does the server respond?
A good developer answers with plain numbers. A defensive reply, or a lecture about how speed doesn't matter much, tells you something about the developer too.
The five-minute test#
Open PageSpeed Insights, a free tool from Google that doesn't need a signup. Enter your URL and look at the mobile result, then test two or three competitors. In five minutes you'll know more about your local market's websites than most sales decks will ever tell you.
The sites I build take this test in public. Each case study on this site with Lighthouse figures links the live build next to the test, so you can run the numbers yourself. I'd ask the same of whoever builds yours: a public score on a live site that you can check.