Skip to content

Services Build and fix

the site takes six seconds to load

Website speed optimization service, measured

from $150

Website speed optimization service for WordPress: images, render blocking, plugin bloat and hosting. Measured before and after, on the hardware visitors use.

What's included

  • Core Web Vitals audit covering LCP, INP, and CLS, on mobile first
  • Images converted, resized, and lazy loaded properly
  • Render blocking CSS and JavaScript untangled
  • Plugin audit: what is loading, on which pages, and what can go
  • Caching and compression configured, not just installed
  • Font loading fixed so text is not invisible on first paint
  • Before and after numbers from field data, not only a lab score

Price

Priced on what is actually slowing the site, which I can tell you before quoting. A theme problem and a plugin problem cost very different amounts to fix.

See every service price

Speed problems are almost never mysterious. They are a hero image nobody compressed, a slider plugin loading a library on every page, a theme shipping CSS for a feature the site does not use, and a font that hides your text for half a second while it downloads.

Why the scores disagree with each other

What makes it feel mysterious is the measuring. A desktop Lighthouse run on a fast machine says the site is fine. Your visitor on a three year old Android over patchy mobile data is having a different experience, and that is the one Google records as field data.

Lab scores are a simulation and field data is a recording. When they disagree, the recording wins, and the recording is the one attached to your site in Search Console. So the first thing I do is stop looking at the number everybody quotes and start looking at the one that came from your actual traffic.

What usually turns out to be the problem

Almost always a short list, and almost never the thing the site owner expected.

Images are first, because a single uncompressed hero can outweigh every script on the page. Then render blocking: CSS and JavaScript the browser has to finish before it will show anything at all. Then plugins loading everywhere to serve one page, which is the most common WordPress failure and the one that gets worse every year the site exists. Then fonts, which are cheap to fix and hide your text while they download. Hosting is on the list, but it is rarely the answer, and upgrading the plan is the most expensive way to not solve this.

How the work runs

Changes go in one at a time with a measurement after each, so it is clear which one paid for itself. A stack of simultaneous changes and one final score tells you nothing, and it is how people end up paying for six fixes when two did the work.

The plugin count usually goes down rather than up. That is the part clients are least prepared for and the part that keeps paying after the invoice, because every plugin removed is one fewer thing to update, break and re-fix later.

Where this sits

The sites I build ship fast by default, because the speed work happens during the build rather than after it. This service is for the ones that did not.

Speed is also a technical SEO problem in disguise on some sites: pages so slow the crawler gives up are a different failure with different symptoms, which is technical SEO rather than this. If the speed work is meant to hold after it is done, maintenance is what stops the next plugin putting it back.

Send me the URL and I will tell you what is actually costing the seconds.

  1. step 1 · measure honestlyLab scores on a fast connection flatter every site. I measure mobile and throttled, and check field data where there is enough traffic to have any, so we fix what your visitors actually feel.
  2. step 2 · find the real costUsually it is a handful of things: an unoptimized hero image, a slider nobody uses, four plugins loading on every page to serve one, and a theme shipping CSS for features you never turned on.
  3. step 3 · fix and re-measureChanges go in one at a time with a measurement after each, so we know which one paid. A stack of simultaneous changes and one final score tells you nothing.

before you ask

Common questions.

Will a caching plugin fix this on its own?

Sometimes, partly. Caching helps time to first byte and repeat views, and it does nothing for a 2MB hero image, a layout that jumps while it loads, or JavaScript that blocks interaction. Most sites already have a caching plugin installed and misconfigured when they come to me.

Does site speed affect rankings?

It is a real but modest ranking factor and a much bigger conversion factor. Do not do this expecting rankings to jump. Do it because a slow site loses people before they see anything, and because Core Web Vitals is one of the few site quality measures Google publishes openly.

Can you do this without breaking my site?

Speed work changes how a site loads, so yes, it can break things when done carelessly. Changes go in one at a time on a site that can be rolled back, and anything I cannot verify visually across the main templates does not ship.

Is this only for WordPress?

WordPress is where most of this work lands, because it is where the plugin and theme bloat lives. The same audit applies to a hand coded or Shopify site, though the fixes look different.

My host says the site is fast. Why does PageSpeed disagree?

Because they are measuring different things and both are telling the truth. A host measures how quickly its server responds, which is genuinely their responsibility and is often excellent. PageSpeed measures how long the visitor waits for something usable, which includes everything the browser has to download and run after the server has finished. A fast server sending a 2MB image and eleven scripts still produces a slow page, and no amount of upgrading the hosting plan touches it.

What score should I be aiming for?

None, and chasing a number is how this work goes wrong. A green Lighthouse score on a desktop run is easy to buy with tricks that hurt real visitors, like deferring things that should load immediately. What is worth watching is the field data in Search Console, because it comes from actual people on actual devices, and whether the page feels usable on a mid-range phone. If both of those are fine, the score is decoration.

Everything else I run.