Get Started

Web Development

WordPress, built to score 90+

The median WordPress site scores 44 on mobile PageSpeed. Ours are published, retested every week, and you can run them yourself.

palmcm.com
A client homepage, built by Itec, rendered in a desktop browser
The same homepage on a phone

Retested every week

Scripts load after paint

Measured

Not a claim. A measurement.

Last run Oct 20

Google PageSpeed Insights (Lighthouse), mobile and desktop. Retested weekly.

Site
Industry
Mobile
Desktop
Tested
Run it
Wealth Management
Mobile
Desktop
Oct 20
Technology
Mobile
Desktop
Oct 20
Mobility Equipment
Mobile
Desktop
Oct 20

Every domain is live. Open PageSpeed Insights and check any of them.

Full scores, Core Web Vitals and history on Results

Lab scores move a few points between runs with server load and network conditions. We publish whatever the weekly run returns.

0–4950–8990–100

The Numbers

Why most WordPress sites don't score like this

WordPress runs over 40% of the web, on every tier of hosting, with any theme and any number of plugins. That's exactly why its average is low. The median WordPress site scores 44 on mobile PageSpeed, and sites built on the two most common page builders, Elementor and Divi, have a median of 40. Closed platforms do better on average: Wix and Webflow sit around 60, because the platform limits what anyone can add.

The gap isn't the platform. It's the default build: a marketplace theme, a page builder, and a long list of plugins, each adding CSS and JavaScript the page never uses, on shared hosting with a caching plugin bolted on to compensate. That stack produces a 44 reliably. A different stack produces a different number, just as reliably.

Mobile PageSpeed scoreThe sites on our board, against the platform medians

Our builds

palmcm.com92
macrotg.com97
slcstairlifts.com90

Platform medians

Webflow Platform62
Wix Platform61
WordPress All sites44
Elementor, Divi WordPress page builders40

Both are Lighthouse mobile performance scores, out of 100: ours from the weekly PageSpeed Insights run, the medians from HTTP Archive's homepage crawl.

WordPress isn't slow. Most WordPress builds are.

Median Lighthouse mobile performance score, HTTP Archive Technology Report, August 2026 crawl. Same data as the Compare page.

How Projects Start

New build. From your design. Or a rebuild.

Three starting points, and the same build underneath every one of them.

New build, from scratch

No design, no existing site. We take it through copy, design, development, and launch, in that order. The content is finalized before design starts, and the design is approved before development begins.

Build from your design

You bring a design file, from your in-house designer or another agency, and we build it into a live, performance-built site. We review the file first and flag anything that needs resolving for performance, responsiveness, or accessibility, so scope and timeline are set against a known deliverable.

Need the design first?

Rebuild or migration

Your current site is on a platform you've outgrown, or its performance has slipped past the point of patching. Your content is carried over and improved, your URLs are redirected so your rankings come with you, and the site is rebuilt underneath. The brand stays. The foundation gets replaced.

See every stage on How We Work

How We Build

Eight decisions that account for the difference

None of this is exotic. It's the accumulation of choices most builds skip because skipping them is faster.

LCP Tagged with the Lighthouse metrics each one moves.

  1. A plugin budget, enforced

    • TBT
    • TTFB

    Every plugin is a dependency: its CSS, its JavaScript, its database queries, its update surface, its security exposure. Our builds run on a fraction of the typical plugin count, because each one has to justify its weight against a performance budget. Most don't.

  2. No page builder tax

    • LCP
    • TBT
    • DOM size

    Conventional page builders output large CSS and JavaScript payloads whatever the page actually uses, plus wrapper markup several layers deeper than necessary. Our builds ship semantic HTML with a shallow DOM. Less markup renders faster, and it's what makes the accessibility and schema work straightforward instead of a fight.

  3. Token-driven CSS, not per-element styling

    • FCP
    • Unused CSS

    Styles come from a small system of design tokens rather than rules generated per element. The stylesheet stays small and cacheable, and it doesn't grow with page count.

  4. Critical CSS first, the rest deferred

    • FCP
    • LCP

    Above-the-fold styles arrive with the page, so first paint doesn't wait on a stylesheet request. Everything else loads without blocking render. On a typical marketing page, this is most of the LCP story.

  5. Caching at the server, not in a plugin

    • TTFB
    • LCP

    Pages are cached at the server and the edge, before WordPress even starts. A caching plugin still has to run PHP before it can serve a cached page. Skipping that step is a large part of why the same site can score very differently on different hosting.

  6. A real image pipeline

    • LCP
    • CLS

    Modern formats, responsive sizes matched to the actual layout, explicit dimensions so nothing shifts, and lazy loading below the fold, with the hero image loaded first. Images are the most common LCP element and the most commonly mishandled.

  7. Font discipline

    • FCP
    • CLS

    Subset to the characters in use, preloaded, and shown immediately with a swap strategy. A few weights, not a dozen. Every additional weight is another request in the critical path.

  8. Third-party scripts, loaded last

    • TBT
    • INP

    Analytics, chat widgets, embeds, and tracking pixels are the most common reason a good score goes bad. We load them after the page has painted (on the visitor's first interaction, or a few seconds after load), use lightweight placeholders for heavy embeds, and tell you what a new tool will cost in load time before it goes in.

    One client's tracking stack was costing them 25+ mobile points; loaded this way, every tracker still fires and the site scores in the 90s.

    64–67Mobile, tracking stack in the page head
    94–97Every tracker still firing, loaded after paint
    Read the case study

This is why we build rather than patch.

Most of these decisions are made in the first week of a project. Retrofitting them into an existing build costs more than starting over, and produces a worse result.

Platform Migrations

Outgrown your current platform?

Every move is the same two jobs: your content, carried over and improved on the way, and a rebuild underneath it. The redirect map is the part that protects your rankings, and it's included every time.

Redirect map · Included in every migration

Webflow to WordPress

The design rebuilt, your content carried over, and a CMS built for publishing at volume.

Comparing your options first? See how they compare

After Launch

A launch score is a snapshot. Keeping it is the work.

No site holds a 95 on its own. Your marketing team adds a scheduling embed. A plugin update ships a new dependency. A 4MB hero image goes up through the CMS. Each one is reasonable in isolation, and together they're why so many sites launch fast and drift.

That drift is the normal condition, not a failure. It's also why our Growth Plans include performance monitoring, and the occasional conversation about whether a new tool is worth what it costs in load time.

Every site on the board above is on an active plan. That's not a coincidence; it's the mechanism.

Mobile score, the year after launchIllustrative, not measured data
  1. A scheduling embed goes in
  2. A plugin update ships a new dependency
  3. A 4MB hero image goes up through the CMS

Built For

Teams that already know what a failing Core Web Vitals report means

Agencies and marketing teams who need a build partner

You handle strategy, campaigns, and the client relationship. You need someone who delivers a technically excellent site on time without you project-managing the development. We work white-label, and your client relationship stays yours.

Marketing managers inheriting an underperforming site

You've seen the PageSpeed report. You've probably already tried an optimization plugin. You want someone who can explain why the number is what it is, and what it actually takes to change it.

Companies where the site is a revenue channel

You're running paid traffic, investing in SEO, or both. Load time sits directly upstream of your cost per lead, and a slow landing page makes every marketing dollar do less.

Results

The full picture behind each score

The board shows where these sites are today. The case studies show what was built and what changed.

palmcm.com homepage, desktop
Design & Development · Growth

Wealth Management

Mobile performance
Accessibility
SEO

Scores publish with the first weekly run after go-live. The pre-rebuild baseline is saved, so this card will show before → after.

Designed, built, and maintained by us end to end. Content and campaigns run by the client's team.

CRO CopyFigma designSchema markup
Read the case study
macrotg.com homepage, desktop
Development · Growth

Technology

Mobile performance
Accessibility
SEO

Scores publish with the first weekly run after go-live. The pre-rebuild baseline is saved, so this card will show before → after.

Built from the client's design. Shows what a marketing tracking stack costs on mobile, with the numbers.

CMPFigma → WordPressHubSpot + tracking
Read the case study
See all results

FAQ

Questions about the build

Can you speed up my existing WordPress site?

Sometimes, but rarely into the 90s. The decisions that produce a high score (build architecture, plugin footprint, CSS strategy, caching) are structural. Optimization on top of a conventional build usually recovers some points without changing what’s underneath. If your site is close and the problem is isolated, like a tracking stack or an oversized hero image, tell us and we’ll say whether it’s worth doing.

What makes your builds different from a normal WordPress site?

Fewer plugins, no page builder overhead, token-driven CSS, critical CSS first, caching at the server instead of in PHP, a real image and font pipeline, and third-party scripts loaded last. The architecture section covers each one. None of it is exotic; it’s the decisions most builds skip.

Will my score stay at 90+?

Not automatically. Third-party scripts, plugin updates, and unoptimized uploads all erode performance over time, and that’s normal rather than a defect. Our Growth Plans exist to manage that drift.

Why WordPress instead of a hosted builder or headless?

Hosted builders do outperform average WordPress on speed, and for a simple brochure site they’re a reasonable choice. Where they stop working is content depth, real integrations, custom functionality, and owning your site. Headless delivers excellent raw performance, but every routine change goes through a developer, and it costs considerably more to run. For companies that need performance and a CMS their marketing team can publish from, a properly built WordPress site is the better total-cost answer. The full comparison is on the Compare page.

Can you build from a design someone else made?

Yes, regularly. We review the file for build feasibility and flag anything that needs resolving for performance, responsiveness, or accessibility before quoting.

Do you work with agencies?

Yes, white-label. Your client relationship stays yours.

How long does a build take?

It depends on where you start. A first build is really two projects: the theme (typography, color, icons, components) has to be worked out and built before anything else, and that first page is the biggest coding stage, so the two together typically run two to three weeks each. Every page after that reuses the theme, and each one is faster than the last. Migrations depend on content volume and the source platform, so we scope them against the actual site.

Check the scores. Then let's talk.

Tell us what you're building. We'll help you optimize it.

Let's Talk ⟶

We respond within one business day.