What kind of sites we build
- Company site. Who you are, what you do, how to reach you. A handful of pages written well beats twenty pages padded out to look substantial.
- Landing page. One page, one goal, built to carry campaign traffic and to measure how many enquiries it actually brings in.
- Multilingual sites. Not machine translation bolted on top: separate versions,
with the
hreflanglinks that tell Google which one to show to whom. - Sites wired to your systems. Catalogue, availability, bookings, forms that land in your business system instead of in an inbox.
- Blog and content. For when the plan is to be found on specific questions, not just on your company name.
What you end up with
- Light pages. No framework loaded to display text. Fewer kilobytes means fewer people leaving on a phone or a slow connection.
- Technical SEO done, not promised. Unique titles and descriptions per page,
structured data, sitemap, reciprocal
hreflang, images with declared dimensions. - Actually readable. Contrast verified above the AA threshold, keyboard navigation, alternative text. It serves people with imperfect eyesight and Google alike.
- Cookie rules respected. A banner that blocks the scripts until someone accepts, not a decorative notice sitting on top of tracking that already started.
- All of it in your name. Domain, hosting, analytics account and source code. If you change supplier one day, you take the site with you.
How we work
- Analysis (free). Who the site is talking to, what it has to make happen, what already exists. By the end you know how many pages you need and what is missing to write them.
- Structure and copy. First the list of pages and what each one says, then the design. Starting from the design means writing everything twice.
- Build. You see the site online at a staging address as it grows, not only at the end.
- Launch and measurement. Going live, submission to the search engines' tools, and one page where you can see visits and enquiries.
What we build it with
We choose after understanding the site, not before. When content changes rarely we build static pages: they load almost instantly, cost very little to keep online, and have no admin panel to keep patched — which is also where nearly every brochure-site breach comes from. When a dynamic part is genuinely needed — a catalogue that updates, a members' area, bookings — we use React and Next.js. If you would rather manage content from a panel, that can be done: it is your call, it carries its own maintenance cost, and we tell you that cost up front.
Two examples you can check
This very site. Static, four languages, one shared stylesheet and one shared script across every page: the home page weighs around 8 KB compressed, fonts aside. That is the kind of result we mean.
Etna Guide Project. Next.js site for a company of alpine and volcanological guides on Mount Etna, rebuilt and put into production in August 2026 with a round of security and performance fixes: the home page went from 68 to 98 on mobile Lighthouse. Four languages, excursions bookable online with confirmation. Go and look.
What it costs
We fix the price up front, for the whole job, and it does not move mid-project. What drives it is always the same list: how many genuinely different pages, how many languages, who writes the copy, whether new photography is needed, and whether the site has to talk to something — a business system, a catalogue, a booking platform. A company site with your own copy in one language sits at the cheap end of that scale; a multilingual site with integrations at the other. The analysis and the quote are free: if the number does not convince you, you have lost one call.
When we tell you no
When all you need is an address and a phone number: a well-kept Google listing costs nothing and gets found first. When the content does not exist and nobody intends to write it — we do not publish a site with placeholder text. When you want an online shop with thousands of items and nobody to keep stock up to date: the problem is not the site, it is who keeps it alive. And when a rebuild is not the answer: if the current site is ugly but works and brings enquiries, you improve the pages that convert instead of throwing it all away.