Website speed and Core Web Vitals
Speed work is where we can show our arithmetic most clearly, so these articles carry the most numbers on the site: median LCP before and after, INP by device class, what a third-party script costs in practice.
The finding that surprises clients most is how much of it comes down to one image being discoverable in the first HTML response, and how little comes from the optimisation plugins they have been paying for.
The twelve rebuilds article is the core of it. The design and ecommerce pieces cover the decisions that create speed problems in the first place.
We measure in the field over 28 days, before and after, segmented by device, because a lab score on office fibre tells you almost nothing about a mid-range Android phone.
SEO best practices for 2026: the checklist we run first
Every audit we start opens with the same eight checks, in the same order. They take a day, cost only attention, and find more revenue than a content plan.
What actually moves Core Web Vitals: twelve rebuilds measured
Twelve client rebuilds, measured over eighteen months. Four changes account for most of the improvement, and the rest was noise. Here are the numbers.
Building an ecommerce site that scales past the first big month
Every store we have rebuilt failed for the same three reasons, none of them the platform. Here are the decisions to make before the platform question.
Web design trends worth keeping in 2026, and three to skip
We judge design trends by what they cost to build, maintain and load, not by how they photograph. Here is what survived that test this year.
One email a month. What we shipped, what broke, what we learned.
No list-building tricks. Unsubscribe is one click and we do not ask why.
Reading is cheaper than rebuilding twice.
Send us the URL. We will tell you which of these articles applies to your site, which one you can ignore, and what we would fix first.