Other · Los Angeles, CA · Built August 2026
Larchmont Digital — our own site
Website designLocal SEO
Results
Lighthouse Performance, mobile
90 — Google's 'good' bar100
Top of the scale
Lighthouse Performance, mobile moved from 90 — Google's 'good' bar to 100, a change of Top of the scale.
Largest Contentful Paint, mobile
2.5 s — Google's 'good' threshold1.5 s
40% under the bar
Largest Contentful Paint, mobile moved from 2.5 s — Google's 'good' threshold to 1.5 s, a change of 40% under the bar.
Cumulative Layout Shift
0.1 — Google's 'good' threshold0
Nothing moves
Cumulative Layout Shift moved from 0.1 — Google's 'good' threshold to 0, a change of Nothing moves.
The problem
Any agency can promise a fast, accessible, privacy-compliant site. Almost none will let you measure theirs. We had to make our own site the argument, because a prospect can check it in thirty seconds and no sales call survives a slow homepage.
What we did
Static-first build with no client-side framework on content pages, self-hosted variable fonts, and images sized at build time. Consent Mode v2 with denied defaults, blocking the Google Tag Manager container itself — not just the tags inside it — until the visitor chooses. Structured data published as a service-area business, with no address we cannot stand behind.
Evidence
Why our own site is the first case study
We sell websites, local SEO and paid ads to businesses that have been burned before. The people who take our call have usually paid an agency that reported impressions going up while the phone stayed quiet. They are not persuaded by adjectives.
So the first thing we ask them to do is check us. Paste larchmontdigital.com into PageSpeed Insights and read the four numbers. If we cannot get our own site right, nothing else we say is worth much.
The numbers, and what they mean for a customer
Performance 100, on a phone. Most people find a local business on a phone, on mobile data, while doing something else. Largest Contentful Paint is when the main thing on screen actually appears: ours is 1.5 seconds on mobile and 0.34 on desktop. Google calls anything under 2.5 seconds good. The bar is not the goal.
Cumulative Layout Shift 0. Nothing on the page jumps while it loads. This is the score that costs businesses money invisibly: a button that moves under a thumb turns an enquiry into a misclick, and the visitor rarely tries twice.
Total Blocking Time 0 ms. No script holds the page hostage before it responds to a tap.
133 KB for the entire page. The median page on the web is several times that. Weight is not vanity — on a weak connection it is the difference between a visit and a bounce.
Accessibility 100. Contrast, focus order, form labels, and every image described. This is a legal exposure for a business with customers in California, and it is also just how you avoid turning away people who use a screen reader or a keyboard.
How it was built
No client-side framework renders the content pages. The pages that need interactivity — the qualification form, the audit tool — load their JavaScript only when the browser is idle, and only on the pages that use them.
Fonts are self-hosted variable files, subset to the characters we actually use, so there is no third-party request blocking the first paint and no flash of unstyled text.
Nothing that measures anything loads before the visitor chooses. That includes the Google Tag Manager container itself, not only the tags inside it, because a container request is already a request to a third party. Consent Mode v2 ships with every category denied by default, and the site checks Global Privacy Control before it shows anything.
What this means if you hire us
Your site will not score 100 in every category — ours can, because we control every byte and we are not running the plugins, chat widgets and booking embeds that a real business needs. That is honest, and it matters.
What carries over is the method: measure before changing anything, ship the smallest thing that moves the number, and never let a marketing tag be the reason a page is slow. The audit we run on your site before we ever speak uses the same measurements you just used on ours.
Marketing results only, published with the client's permission. Results vary by business, market and budget, and Larchmont Digital makes no claim that they are typical or repeatable.