Web technology is often presented as a matter for specialists, reserved for people who know its vocabulary. Yet its principles come down to a few notions. Knowing them makes it possible to judge what a provider is offering, to keep a website from ageing fast, depending on a single person or losing its videos, and sometimes to do yourself what you thought was out of reach.
Hosting, domain name, content management software, backup, search engine optimisation, video, social networks: I offer here an overview of these basic notions, written for people who are not computer specialists. I would like to give them what they need to make their own choices, talk things through with a provider, and do part of the work themselves.
Websites that age fast and depend on a provider
Many websites of associations, cultural venues or small businesses age very quickly. After four or five years, they are rebuilt from scratch, and part of the content is often lost along the way. Only the provider knows how to change them. The domain name was registered in the provider’s name. Nobody in the team knows where the site is hosted, or whether it is backed up. And the day the provider goes out of business, the site stays as it is, without updates, until it stops working.
Why? Providers do what they can, with the specifications they are given. But the person who commissions the site often lacks the knowledge that would allow them to judge what they are offered, to ask the right questions, to request the right guarantees. They were told it was technical, and therefore a matter for specialists, and that is where the misunderstanding lies: the principles of the web are simple, and jargon often serves to hide that simplicity. A few notions are enough to read a quote, understand what the choice of a tool implies, or notice that a promise means nothing. “Thanks to us, you will rank well on Google” is one of them: you never “rank well” in the absolute, only for specific searches.
And once these notions are acquired, you discover that you can do many things yourself, up to installing software or changing the look of a site.
Five words to follow a technical discussion
The web rests on computers linked to one another by the internet, some of which, the servers, keep pages and send them to those who request them. The following five words come up in any discussion with a provider:
- The server is a computer that stays on permanently, at a hosting company. It stores the site’s files and sends them to the people who visit it. The hosting company rents space on this server, which is often shared among many sites: this is called shared hosting.
- The domain name is the address of the site, such as myassociation.fr. It is registered for one or more years with a registrar, and points to the server that hosts the site. For addresses ending in .fr, the registry is run by Afnic.
- A web page is a text file, written in HTML. Tags indicate in it what is a heading, a paragraph, a link or an image. HTML belongs to no one: CERN put the web software into the public domain in April 1993, and anyone can write a page without asking for permission or paying for a licence.
- The browser (Firefox, Chrome, Safari, etc.) receives this text and displays it. You can see the code of any page with the shortcut Ctrl+U (Cmd+Option+U on a Mac), and so check what it really contains.
- The content management system, or CMS, builds pages from a database. WordPress, SPIP, Joomla, Drupal, etc.: you write your articles in an interface, without touching the code, and the software lays them out according to templates. According to W3Techs, more than 40% of the world’s websites are built with WordPress.
Who owns your website?
There are two main ways of having a website. On an online platform such as Wix or WordPress.com, the company provides both the tool and the hosting: you sign up, choose a template, and publish. It is quick, but the organisation that makes this choice depends on a company whose prices and rules it does not set, and which may disappear; it is often hard to leave with your content. The other way is to install free software yourself, at a hosting company, such as WordPress (the software, not to be confused with the WordPress.com platform) or SPIP, for around a hundred euros a year; you can then move whenever you decide, with all your content, provided you take care of it. I summed up this choice in one sentence during a webinar for the Fondation Afnic in October 2025: “independence comes at the price of responsibility”.
Before signing with a provider, or to take stock of an existing site, I suggest you ask yourself these questions, or ask them of your provider:
- Who is the holder of the domain name, and who receives the renewal notices?
- With which hosting company is the site installed, and who holds the login details?
- Which software is the site built with (free or proprietary), and who installs its security updates?
- Where are the videos stored: on the site’s server, or on a platform?
- How do you get your content back (texts, images, videos) if you change provider?
- Who in the team will be able to publish and correct a page, and who will train them to do so?
- Who makes the backups, how often, and where are they kept?
If the answers boil down to “we take care of it”, the site in practice belongs to the provider, and it is time to discuss it with them.
Durability: a website is designed for ten years
To my mind, a website is first and foremost an archive. Each published page keeps its address and can be found years later by someone looking for exactly what it contains. The site still has to last, and I see three conditions for that.
First, backups. During the night of 9 to 10 March 2021, a fire in an OVHcloud data centre in Strasbourg took some 3.6 million websites offline, and some organisations, having no copy elsewhere, lost theirs. A website is made up of files (videos included, when they are on the site) and a database; you copy them regularly, keep a copy somewhere other than at the hosting company, and check at least once that you know how to put them back online.
Next, updates. A CMS is software installed on a server permanently connected to the internet, and malicious people look for security holes in it. The teams that develop it publish fixes, which have to be installed. A site that is never updated is at risk of being hacked, or of ceasing to work.
Finally, addresses. When a page changes address during a redesign, all the links that led to it, on other sites, in emails, in search engine results, are broken. The old addresses are then redirected to the new ones, and old pages are kept online, even when what they announced is over. In Liability and computer backup and in Our fragile memory, I developed what this fragility puts at stake for our collective memory.
How a page gets found
A search engine such as Google works in three stages. Programs called robots browse the web by following links from one page to another. The text of the pages collected is stored in an immense index. When someone types a search, the engine ranks the pages of the index that match it, taking into account in particular the words on the page, its headings, and the links that other sites make to it.
A few practical consequences follow. A page appears in the results of specific searches, made with the words used by the people searching: a stamp-collecting association has little chance of appearing for “stamp collecting”, much more for “stamp collecting” followed by the name of its town. Text matters most: an explicit title, subheadings, a first paragraph that says what it is about, images described by a text. And search engines are no longer the only ones collecting the content of websites. When you ask an artificial intelligence assistant a question, the answer is written from pages collected on the web, and an organisation that publishes clear, dated and up-to-date pages has a better chance of appearing in it.
Hosting your videos like your images
I have long argued that videos should be treated like images, and the technology allows it. A video is a sequence of images and a sound, compressed by a coding process (the most widespread on the web is called H.264). It is much heavier than a photo, but it goes online in the same way: you put the file on the site’s server, and the page displays it. The HTML <video> tag, which every browser reads, does for a video what the <img> tag does for an image, and in SPIP as in WordPress, you attach a video to an article just as you attach a photo. The file simply has to be prepared for the web. In Best practices for putting videos online, I gave a benchmark: a file of ten to twenty megabytes per minute plays more or less everywhere. The free software HandBrake lets you make this file yourself, from the file exported by an editing program, by ticking the “Web Optimized” option.
Yet most organisations put their videos on YouTube or Vimeo, often on the advice of their provider, even though nothing technical obliges them to. The videos are then stored at these companies, not on the site: when you back up the site, you back up the texts and images, but not the videos. Should the platform change its terms, close an account or go out of business, the site’s pages show nothing but empty players. By uploading your videos, you also accept terms of use that grant the company rights over them. YouTube, finally, adds advertising and suggests, at the end, videos other than yours.
An organisation’s videos are part of its heritage, just like its texts and photos. In Les Lilas, where I worked for four years with the town’s cultural affairs department, we built together a video platform that hosts on its own server the films made by residents. The files remain at hand, and they can for example be screened before the main feature in the town’s cinema. The work of putting them online also made it possible to recover films that would otherwise probably have been lost, or never finished. The story of this project is told in Video platform « The digital factory in Les Lilas ».
Nothing then prevents you from also publishing a video on YouTube, when you want it to circulate widely at a given moment; the reference version remains the one on the site. And a video is only found by search engines through the text that accompanies it: a title, a description, and if possible a transcript or subtitles, which also serve deaf and hard-of-hearing people. A video placed on a page of the site, surrounded by this text, therefore has a better chance of being found than a video posted on a little-visited YouTube channel, as I showed for cultural venues in Cultural venues and social networks.
Social networks, to make yourself known
Facebook, Instagram, LinkedIn or TikTok belong to private companies. You publish on them in a feed, where each post is quickly buried by the ones that follow, within a few hours or a few weeks. What followers see depends on the ranking set by the platform. Texts there are short (2,200 characters at most for an Instagram caption). Many posts are only displayed to people logged in to an account, and they are poorly indexed by search engines; since July 2025, Instagram has let search engines index some public photos and videos, but it is the company that decided this, and it can go back on that choice.
Social networks are therefore useful for making the organisation known, announcing its events, relaying, and talking with the public. The site, for its part, is the place for reference information, for the history and the stories of the organisation. I suggest writing first on the site, then relaying the page on social networks, with a link. I explain this division in Editorial strategy for an organisation: the website as a living archive and reference source, and social networks to make it known.
What you can do yourself
With these notions, a good part of the day-to-day work on a website is within the reach of any team:
- publishing an article, correcting opening hours, adding an image in the CMS;
- displaying a page’s source code (Ctrl+U) and checking its title, its subheadings, the descriptions of its images;
- reducing the size of a photo, or encoding a video for the web and putting it on the site like an image;
- downloading a backup of the site from the hosting company’s interface;
- reading the visitor statistics to see which pages are viewed;
- understanding a quote, and asking why a line costs what it costs.
To go further (installing software, changing a template, writing a few lines of code), some support is often enough, and today you can have code explained to you by an artificial intelligence, as I recounted in To Code or Not to Code ?. I come back to each of these notions in more detail in the other articles of this section. I presented a first version of this overview during the three webinars I ran for the Fondation Afnic in October 2025, whose recordings are gathered in Web seminar resources : « better communication on the web ».
