We put two websites on a scale. Here's what 2.6 megabytes actually costs you

Before we ask anybody to hire us, we figured we should prove we can do the thing we sell. So we did what we always do: we measured. We took a page we built and a competitor site we measured — a real, live, professionally-made site in our own industry, the kind that looks good and costs real money — and we ran both through the same instruments. Not opinions. Instruments.

This post starts a series. Every month or so, we'll measure a real local business website the exact same way we measure our own client work, strip the name off it, and publish what we found. Nobody gets embarrassed. Everybody gets to learn. And every single thing we publish, you can check yourself.

Here's the first one.

The numbers, straight

That's a 61-to-1 difference in weight. Their styling code alone — just the styling — is 37 times bigger than our whole page.

Now, two honest caveats before you quote any of this, because we print our caveats instead of burying them:

First: our page was a prototype — a hero and a services section, no photo gallery. Their site carries real photographs of real finished work, and real photography is heavy. It's also worth it — a photo of an actual job sells better than any paragraph ever will. A full site of ours with a real portfolio would weigh more than 44 KB, and it should. The honest comparison isn't the total; it's everything else on that table. Their 1,668 KB of styling code buys nothing a customer can see.

Second: we didn't measure their site's speed score, so we're not claiming one. We measured what we could measure cleanly and left the rest blank. A blank cell is honest. A guessed one isn't.

Okay — but what does a kilobyte cost?

Here's the part that matters if you run a business out here.

Every person who visits your website downloads it. All of it. The whole 2.6 megabytes, every visit, before the page settles down and behaves. In town, on good fiber, that's a pause. But your customers aren't always in town on good fiber. Around Cedar Creek Lake, they're on a phone, on rural LTE, standing in the backyard looking at the problem they want to hire you for. Out there, a heavy page isn't a pause. It's a blank screen with a spinner — and a back button.

And the back button is the whole story. Nobody emails you to say your website was slow. They just leave, and they call the next name on Google, and you never find out it happened. A slow site doesn't fail loudly. It fails silently, all day, and the only evidence is jobs that quietly went somewhere else.

We won't dress that up with a made-up statistic — you've seen the ones floating around, "every second costs you X percent." We didn't run that study, so we won't quote it. The plain logic is enough: the person searching for your trade at 9 p.m. gives your site a few seconds. Either it shows up, or you don't exist.

Where the weight comes from (it's not your photos)

Here's what surprises most owners: the bloat usually isn't the pictures. On the site we measured, the single biggest load was the styling code — 1,668 KB of it, across 34 stylesheets. Why so much? Because most sites are built on page builders, and a page builder ships every style it could ever use, whether your page uses it or not. It's like pouring a foundation for a hotel and building a bait shop on it. Nobody decided to do that. It's just what the tool does when nobody's watching the scale.

That's also why this doesn't fix itself as sites get bigger. A hand-built page grows by what you add. A page-builder page starts fat before you've added anything.

The 22 scripts tell the same story. Every plugin, every popup, every tracking pixel somebody added in 2021 and forgot — each one is another thing your customer's phone has to download, parse, and run before the phone number becomes tappable. We got our page to two scripts by asking one question about every single piece: does this earn its bytes? Most things don't.

What their site does better — because a real teardown says so

The site we measured wins something important, and we'd be lying by omission if we skipped it: their photography is better than ours, because they have real photographs of real finished work and our prototype had none. Their pictures do real selling. That's not a small thing — it might be the biggest thing. A fast page with nothing to show is a clean, empty showroom.

That's the standard this series holds itself to. Every teardown will tell you what the site gets right, because a measurement that only finds what it wanted to find isn't a measurement.

Check your own site in the next five minutes

You don't need us for this part:

1. Open your site on your phone, off wifi, outside town if you can. Count the seconds honestly.
2. Search "PageSpeed Insights," paste in your address, and look at the mobile score. It's free, it's Google's own instrument, and it's the same family of tools we used above.
3. Look at the page-weight number. Under 1 MB is disciplined. Over 2.5 MB and your customers on lake internet are paying a tax you can't see.

If the numbers are ugly, your web company isn't necessarily bad — most of the time, nobody ever measured. That's the quiet truth of this whole series: almost nobody measures. We do, it's most of what makes us different, and it's why we publish the instruments instead of adjectives.

Want the same measurement run on your business — site, Google profile, reviews, the works — with the findings yours to keep whether you ever hire us or not? That's the free audit. We send the good verdicts too.