Answer in brief
Asked about pages, most people answer with menu items. But a site has three layers, and the third — the utility pages — gets built in a rush or not at all. The full list by layer.
A page list is not a menu structure
Asked what pages a website should have, most people answer by listing menu items: home, about, services, contact. That is not a list of pages but a list of what is visible in the header.
A real site has three layers. The first is what sits in the menu. The second is pages reached by links from inside: individual services, articles, case studies. The third is utility pages that are in neither the menu nor the links, and are still needed.
The third layer is almost always the forgotten one. It gets built in a rush before launch, or not at all, and then surfaces at the least convenient moment.
Below is the full list by layer, with what each page is for, and a check that shows what you are missing.
Four pages you cannot do without
The minimum below which a site stops being a working tool.
The home page. Its job is not to describe the company but to answer in five seconds what you do and who for, and to move somebody into the right section.
A page for what you sell. One for each service or product, not one general page covering everything. More on that below — this is where corners get cut most often.
Contact. Not a form and not one phone number, but a proper page, whose contents deserve their own discussion.
And a confirmation page after a form is sent. It is skipped more than any other, and it is why enquiries arrive twice or not at all.
Contact: what belongs on it
A page built in five minutes that works as proof the company exists.
A phone number as a link rather than plain text, so it dials on a tap from a phone. An email address as a link too.
An address, if there is a physical place, with a map and a spoken landmark. A map does not replace a description: second floor, entrance from the courtyard saves people fifteen minutes.
Opening hours and the time it takes to answer enquiries. The second matters more than the first: people want to know when somebody will get back to them, not when you break for lunch.
And company registration details if you work with organisations. For many companies their absence is a reason not to start the conversation.
About: what it is actually for
This page is read less often than its authors imagine, and what it decides is not what usually gets written on it.
It is not about a history starting from the founding year. It answers one question: can you be trusted with money and time.
What works here is specific: names of people and their roles, photographs of a real place, names of companies you worked with, figures that can be checked.
What does not work: years in the market, team size, the words professionalism and individual approach. Everyone writes those, including people who cannot do the job, so they do not distinguish.
A practical test: take any claim from this page and ask whether a competitor could have written it. If so, remove it.
Service pages: one per service
This is where corners are cut most often, and it is the most expensive saving on a site.
One general services page listing twenty items answers no specific query. Somebody looking for one thing gets a catalogue and an invitation to work it out themselves.
A page of its own belongs to every service that has its own question and its own buyer. If two items differ only in name, that is one page, not two.
Our catalogue holds a hundred entries, each with its own page, text, price and answers. That is exactly what makes it possible to answer a hundred different queries instead of one.
The opposite extreme exists too: twenty pages written about the same thing in different words start competing with each other. A separate page is warranted when it has a question of its own.
Prices: a separate page or on each service
A contested question with a practical answer: both, but differently.
Every service page should carry its price or a range. Somebody who arrived for a specific service should not have to leave and hunt for a price list elsewhere.
A separate pricing page is needed too, because people search for it as its own query. It gives an overview and lets items be compared against each other.
The critical thing is that they must not diverge. A price list updated in one place and forgotten in another does more harm than having neither page.
And if an exact price is impossible, give the order of magnitude and what moves it. An absent price reads as risk rather than flexibility.
Utility pages people forget
The third layer, the one built in a rush.
Form submission confirmation. Its own address rather than a popup message: that way you can count enquiries in analytics and somebody sees the confirmation even on a slow connection.
The error page. That gets its own section below.
The site search results page, if there is a search. It is often left unstyled, and the people who land on it are exactly those who already failed to find what they needed in the menu.
Delivery and payment terms, if you sell anything. This is not a legal page but a practical one: timescales, zones, methods, what happens on a return.
And a page for people arriving on an old address, if the site had a previous version. After our own move there were six hundred and seventy-three such addresses.
The 404 page
The most underrated page on a site: the people who land on it have almost left already.
What belongs on it: a clear message without technical jargon, a link to the home page, links to two or three main sections, and site search if it exists.
What does not: an automatic redirect to the home page. The visitor does not understand what happened and concludes the site is broken. Search engines, incidentally, reach the same conclusion.
And something technically important: the error page must return a 404 status rather than a 200. A handsome page that formally answers everything is fine creates thousands of empty pages in search.
It takes a minute to check: open any non-existent address on your site and look at both what is shown and what the response header returns.
Legal pages
A section where I speak about structure rather than law: requirements differ by country, and the exact wording is a question for a lawyer.
In practice a site usually needs: a privacy policy, consent at the point of form submission, and, if you sell anything, published terms of sale.
What matters technically: the link to the policy belongs beside every form, not only in the footer. People consent at the moment of sending, and the text has to be available exactly there.
Check separately that the policy lists what you actually collect. If you have added session recordings or analytics tags and the policy says nothing about them, the document does not describe what is happening.
And keep these pages open to indexing. There is no reason to close them, and their presence signals that the company is real.
What should not be there
Three kinds of page that appear often and make things worse.
Empty under construction pages. A section promising that something will appear soon reads as a notice that nobody is looking after the site. Better not to publish until it is ready.
Duplicates of one service on different addresses. They appear on their own as pages get added for advertising, and afterwards those pages compete with each other in search.
And pages nothing links to. If a page is needed, links should lead to it. If there is nowhere to put a link, the page is probably not needed. That is how we discovered an item passport lookup page with nothing pointing at it from any product page.
How to check for completeness
Five actions to find out what you are missing.
Write down the tasks people arrive with and confirm each has a page where it gets done.
Walk it backwards: every page should have a task. A page without one is either surplus or you have found a task you never wrote down.
Submit a form and check the confirmation arrives on its own address.
Open a non-existent address and look at both the error page and the response code.
And check whether at least one link leads to each utility page. A page that exists but cannot be reached does not exist for a person.
Practical checklist
- Write down visitor tasks and find a page for each of them.
- Walk it backwards: every page should have a task of its own.
- Check the form confirmation has its own address.
- Open a non-existent address and check the response code, not just the page.
- Put the policy link beside every form.
- Reconcile prices on service pages against the separate pricing page.
- Confirm at least one link leads to each utility page.
Questions and answers
Which pages are mandatory on any website?
Four: the home page, a page for what you sell, contact, and a confirmation page after a form is sent. The last is skipped most often, and it is exactly why enquiries arrive twice or not at all.
What belongs on a 404 page?
A clear message without jargon, a link to the home page, links to two or three main sections, and site search. What does not belong is an automatic redirect to the home page. And the page must return a 404 status rather than a 200.
Does every service need its own page?
Yes, if the service has its own question and its own buyer. A general page listing twenty items answers no specific query. But if two items differ only in name, that is one page, not two.
Prices on each service or on a separate page?
Both. A service page should carry its price or a range so nobody has to leave and hunt for a list. A separate pricing page is needed because people search for it as its own query. The critical thing is that the two must not diverge.
Where should the privacy policy link go?
Beside every form, not only in the footer. Consent is given at the moment of sending, and the text has to be available right there. While you are at it, check the policy lists what you actually collect, including analytics tags and session recordings.