What is a website — and the one axis that orders every type

„What kind of website do we need" sounds like a question about categories. It is a question about one variable: what the site has to accept. Everything else — page count, technology, looks, price — follows from that answer rather than causing it.

This section sorts out the basics and points you to the article that goes into detail. None of them is explained from scratch here. What you will find here instead are two things that belong to a switchboard and to no single article: what you are actually buying when you buy "a website", and how the types sit relative to one another.

What you buy when you buy "a website"

Four things, paid for separately, and none of them is the website. Worth separating before the first number lands, because two quotes that differ by a factor of two usually differ in how many of these four they cover.

The address — the domain. The name people find you under, rented by the year and renewed. It is not owned outright and it is not part of the site — it outlives every version of it and every contractor. It is the only one of the four whose loss is irreversible, which is why the one question to ask before signing is: whose name is it registered in.

The place — hosting. The server the files sit on and are served from. Paid on a cycle whether or not anyone visits. We break that down in the cost section.

The content. Text, photographs, prices, service descriptions. The largest item you control and the most common cause of delay — a contractor can edit it but cannot invent it. It has a whole section of its own: content and media.

The software. What turns content into a site and lets you change it without a developer. The widest choice of the four and the least final — it is the only one you can swap later, though not for free. Tools and technologies are two separate sections.

The practical conclusion: "what does a website cost" is a question about four different bills, three of which return every year. A quote that gives one number without splitting it across those four is not cheaper — it is less legible.

The second conclusion is about what breaks. Each of the four breaks differently and at a different moment, and businesses usually reach us when the one nobody was thinking about has broken:

• A domain registered to the contractor. Surfaces only when you change contractors, or when one disappears. The site keeps working and you can do nothing with it. Checking is free and takes a minute — registrant details are in the public register.
• Hosting that will not let you change the software version. Surfaces at the first forced update, and means that modernising the site is secondary to changing the supplier.
• Content nobody wrote. Surfaces in week three, when the project stalls waiting for service descriptions. It is the most common cause of a missed deadline we know of, and the only one you can head off before signing.
• Software nobody in the company can operate. Surfaces after six months, when "you can edit it yourself, in theory" turns out to mean every change goes to the contractor by email and gets billed.

These four are worth settling before you pick a type, because they apply to every type equally. Only then does the question this section is actually about start to mean anything.

Website types are one axis, not six categories

Most lists of website types read: brochure, business, shop, landing page, blog, portal. A list implies you pick one of six equivalent things. That is not how it works — four of them sit on a single axis, and two are not points on it at all.

Nothing — the brochure site. It confirms you exist and gives a number. The sale happens in conversation; the site's job is to lead to that conversation and make you credible on the way. When that is enough and what has to be on it: brochure website.

Contact — the business website. It takes an enquiry and answers the questions that come up every week in conversation. Each such question is a page, and each page answers for you while you sleep. The usual choice for service businesses: business website.

An order — the shop. It takes the order and the payment. That is the threshold where the project changes category in cost, in timeline and in what maintaining it involves — which is why it has a section of its own rather than a paragraph here.

An account — the portal. Users log in and come back for their data. This is a system, not a site, and one question settles it: does the user have an account?

Now the two that are not on the axis, though lists tend to name them alongside the other four.

A landing page is not a separate type — it is a business website narrowed to a single action, usually for one campaign. It can exist alongside any of the four above, and usually does. What it consists of and what actually decides its outcome: landing page.

A blog or content hub is not one either — it is a content layer you can add to any of the four types. A brochure site with a blog is still a brochure site; a shop with a blog is still a shop. Whether a blog makes sense for you at all is settled in the content section.

Why this distinction deserves its own section: treating the axis as a ladder produces two expensive mistakes at once. The first is adding pages nobody reads, on the belief that a company "outgrows" a brochure site — a service business can stay on one for ten years and be right, not negligent. The second is deferring a system that was needed on day one — a start-up that needs a customer panel does not grow into a portal, it starts as one.

Worth knowing too what changes between neighbouring points, because that is where the difference in price and timeline lives — not in the page count.

From "nothing" to "contact" you add content and URL structure, not functionality. The project grows by however many questions repeat, and it is the only step on this axis you can take in stages without rewriting the rest.

From "contact" to "an order" you add payment, delivery, returns and obligations that did not exist before — terms, settlement, complaint handling. That is a change of category, not of size: what changes is the cost of running it, the response time on failure, and what happens when the site is down for an hour.

From "an order" to "an account" you add something that does not appear in the quote: data you have to hold and protect. A user who logs in leaves their history with you, and from that moment the conversation about security and backups stops being optional.

What every one of them has to do, whatever the type

Six things common to all four points on the axis. Worth knowing before you talk to a contractor, because none of them is an extra and every one of them gets quoted as an extra.

Its own address. Not a contractor's subdomain and not a profile on someone else's platform. It is the only item here whose absence is irreversible; everything else can be fixed later.

Working on a phone. Not in a desktop preview — on your own phone, on mobile data: a form you can fill without zooming, a menu that opens, images that do not push the text off the screen. The test takes three minutes and catches more than an hour of looking at the design.

Contact visible without hunting. A number or a form within one glance, on every page. It sounds trivial and it is the single most common defect we find in audits.

An encrypted connection. A standard now rather than a nicety — browsers enforce it harder every year, and a site without a certificate can simply stop opening. Deadlines and detail: the security section.

The ability to edit it yourself. The content at least: prices, hours, phone number, a service description. If changing a phone number needs an email to the contractor, that is a cost that returns a dozen times a year.

A backup somebody has actually restored. "We have backups" is not an answer. The question is who restored one last and how long it took. We expand on that in maintenance.

Note what is not on that list: animation, scroll effects, background video, graphic trends. Not because they are bad — because they are a choice, and these six are a condition.

Three questions that settle the choice

None of them is about page count or budget. Answering these three is enough to place you on the axis.

Where the sale happens. In conversation — then the site has to lead to it, and more pages will not speed that up. On the site — then you need a basket or an order form, and that is a different category of project.

How many questions repeat. Count the questions that come up in conversation more than three times a month. That is your page list, and it is the only honest criterion for page count we know. Zero repeating questions is a brochure site; ten is a business website.

Whether the site has to remember anything. An order, a booking, a service request, a history. If yes, you move to the right-hand end of the axis and the conversation stops being about a website.

An example, because in the abstract these sound theoretical. A building firm: the sale happens in conversation after someone has looked at the site of the work, so the first question points at the left-hand end. But four questions come up every week — do you do bathrooms, what is the lead time, do you invoice, how far do you travel — so the second question moves the project one point right. Nothing has to be remembered, so the third changes nothing. Result: a business website with four service pages, not a brochure site and not a twenty-page affair.

The same three questions put to a training company give a different answer, though the sector is just as service-based: sign-ups for a date have to be taken and remembered, so the third question moves the project straight to the right-hand end — and the conversation stops being about a website at all.

There is a question zero, asked before those three: whether the site will pay for itself at all. It does not always, and there are situations where we advise against starting with one. The arithmetic and an honest list of those situations: why your business needs a website.

Four ways into this section

Four articles, one per situation businesses arrive with. Named by the situation rather than the topic.

"We are not sure it pays for itself." Arithmetic instead of adjectives: what a site actually takes over, how to tell it has paid for itself, and when it is not the first expense. → Why your business needs a website

"We want the smallest sensible site." The five things that have to be on it, the three conditions under which a brochure site is enough, and the signals that you have outgrown it. → Brochure website

"We have several services and do not know what we need." Four types of business website described by their job, three questions that settle the choice, and the one thing you cannot add later. → Business website

"We are running a campaign and need a page for it." The anatomy of a landing page, what not to put on one, and how it differs from an offer page and from a one-pager. → Landing page

Four sentences we hear most often

Each sounds reasonable and each leads somewhere other than intended.

"The site has to be modern." Modernity is not a criterion, because it can be neither verified nor quoted — and a project commissioned against it gets judged on impression, the most expensive currency in this trade. The criteria you can check are four: does it work on a phone, can someone in the company add a page without a developer, what software version is it on, and can you send an enquiry from it in three minutes. All four take fifteen minutes.

"More pages are better for Google." What counts is not the number but whether anyone is searching for what is on them. Twenty pages restating one topic in different words perform worse than five answering five real questions — one of the rare cases where more work yields a worse result. Which is why the criterion for page count is the list of repeating questions, not ambition.

"We will do a simple one first and expand later." True at one point on the axis and false at two. From a brochure site to a business website you genuinely can go in stages, because what grows is content rather than functionality. From a business website to a shop or a portal you cannot, because the category changes rather than the size; bolting a basket onto a site designed without one usually costs more than building it with one from the start.

"The website will handle our marketing." A website is where marketing leads — it is not marketing. A site nobody visits is a site nobody visits, however well it is built. Where traffic comes from and how long it takes is settled in its own section, and it is the most common reason behind "we built a site and nothing came of it".

What this section does not cover

The basics answer "whether and which kind". Everything after that question has a section of its own — deliberately, because keeping those answers in one place produces a document that is out of date in half of it.

• What it costs and what the bill is made of — cost. No figures here on purpose: they are kept in one place so they do not go stale in five.
• What happens in what order — process, including the order in which changing your mind is still free.
• What gets decided before any code exists — design.
• What to write and where to get photographs — content and media.
• What to build it on — tools and technologies.
• Where traffic comes from — marketing and SEO. The largest part of the answer to "why did this not pay off" lives there, not in the design of the site.
• What can go wrong — security.
• What has to happen monthly — maintenance.
• What to do with a site you already have — strategy.

Selling goods online is the one item on the axis without a section linked here, because it is a different project with different fixed costs and different risks — and it is covered separately rather than as a paragraph in a piece about websites.

What order to read this in

Three routes, depending on where you are.

We have nothing and do not know whether it is worth it. Start with the arithmetic, then brochure website, then cost. Three articles and the decision is made — or deliberately postponed, which is also a decision.

We know we want one, we do not know which. Start with types of business website, then process, then cost. In that order, because a quote produced before the scope is settled is a quote for something nobody has defined yet.

We have a site and something is wrong with it. The basics are not your first stop — go to strategy, which starts by identifying the symptom rather than by deciding to build. The most common mistake at this stage is commissioning a new site before anyone has established what exactly the old one fails at; the new one often turns out equally handsome and equally ineffective, because nobody changed what was actually deciding the outcome.

One note common to all three routes: none of these articles substitutes for counting your own numbers. How many questions repeat, how many enquiries arrive today and from where, what an average customer is worth — four figures you will not find in any article and without which every recommendation is guesswork. All four are already in your business and collecting them takes an hour.