What should a corporate website look like: section by section.
A corporate website should answer, in a few pages, the four questions of a visitor who does not know the company: who is this company, what does it do, why is it trustworthy and how do I reach it. The same answers should also be written in a form that search engines and AI answer engines can read; that is why every section of the site has two readers: the human and the machine.
A corporate site is often thought of as an extended business card. Yet today someone researching a company for the first time usually enters the site not through the home page but from a search result, in the middle of a service page; and sometimes, without opening the site at all, they read the sentence an answer engine took from your site. That is why every section has to stand on its own.
Below we take the sections one by one, separating in each what the human reader is looking for from what the machine reader reads. We use this distinction as the starting point of design in our corporate web design projects. The general checklist is in the guide what to watch for when having a website built.
Who are the two readers of a corporate website?
The human reader looks at the page: they see the heading, the image, the text, the button, and decide within a few seconds whether to stay. The machine reader reads the page’s source: the title tag, the description, the heading hierarchy, the links and the structured data. Google’s own guidelines ask for both together: content written for people, and content the search engine can understand.
The two readers often want the same thing but read it from different places. A slogan set in large letters on the home page tells a human something; if there is no single sentence saying what the company does, it tells the machine nothing.
What should the home page do?
For the human reader, the home page’s job is to give direction: on the first screen, what the company does and for whom should be readable in one sentence, and the next step (services, work, contact) should be one click away. When the home page tries to say everything, it ends up saying nothing.
- For the human → on the first screen, what you do and for whom; a path to evidence (work, references); a path to contact.
- For the machine → a title tag and description stating the company’s name and business; structured data defining who the company is; the site name Google will show in search results.
What should service and product pages look like?
Every service or product group should have its own page and its own address. Five services listed one below the other on a single “Our services” page make all five look weak in search; each one answers a separate question and deserves a separate page.
A good service page starts with the customer’s question and states its claim together with its evidence. In the Karcem case, a measurable quality claim sits at the center of the site — color accuracy — and the certificates are presented in a separate layer as evidence of that claim. On Fersan Tekstil’s seven-language site, every market reads about its products in its own language; that requires a separate address for each language and tags linking the language versions to one another.
- For the human → what you do, for whom, how you do it and evidence showing it; frequently asked questions; the next step.
- For the machine → a single main heading (H1) and a logical order of subheadings; structured data describing the service; links to related guides and case studies.
What belongs on the about or imprint page?
The imprint is the page of trust. Here the human reader looks for the people behind the company, where and since when it has been working, and what it can prove. The machine reader wants to recognize the company as an entity: its name, address, founder and official accounts on other platforms. We covered these links in detail in the guide sameAs and digital identity.
Errors in this layer are silent. For this site’s imprint page, Search Console’s profile page report treated a last-modified date written only as a day (without a time) as invalid; the fix was a date-time with a time zone. The page looked flawless to the human eye — only the machine reader could see the error.
What should the contact page look like?
The contact page’s job is not to lose a visitor who has already decided. The form should be short, required fields should be clearly marked, and when an error occurs the visitor should understand what to fix — including a visitor using a screen reader. How to write error messages is covered in the guide form error messages.
- For the human → email, phone, address or service area; a short form; a sentence saying what happens after the form.
- For the machine → the company’s contact details written the same way in the structured data; labeled form fields.
Does a corporate site need a guides or blog section?
It is not mandatory, but this section often determines the site’s place in search. The home page and service pages answer the question “what do you do”; the “how”, “why” and “which one” questions a customer asks before deciding need separate pages. The value of these pages lies not in their number but in their depth. On 22 August 2026, 13 pages of this site had the status “Crawled – currently not indexed” in Google; the diagnosis was thin content — that day none of the 58 pages exceeded 800 words. Our answer was to write every guide page deep enough to genuinely answer a question; a guide under 800 words now turns the test suite red.
- For the human → evidenced, dated answers to the questions a customer asks before deciding.
- For the machine → each article’s own address, author, publication and update date; an introduction that answers the question in its first sentence.
How many languages should a corporate site have?
As many as the languages your customers search in. A second language is not a translation button added to the corner of the menu but a second site with separate addresses for each language: pages are paired with tags linking the language versions to one another, forms and error messages are translated too, and the structured data is written separately in each language. A language version produced by a plugin that translates text on the fly in the browser usually has no address of its own; a search engine cannot see it as a separate page. For an exporting company, this means a foreign buyer cannot find the site when searching in their own language.
Where should the legal pages sit?
The privacy and personal data notice, the cookie information and the company’s identifying details should be reachable from every page of the site, usually from the footer. Which texts are required in which situation, and their official sources, are covered in the guide what a website must have; the information there is not legal advice either.
What ties the sections together?
The menu, the breadcrumb trail and in-page links turn the sections into a single site. For the human, this means knowing where they are and where they can go; for the machine, it means understanding the relationships between pages. On this site Search Console found a breadcrumb trail with an unnamed link; since then the site’s generator stops the build when a breadcrumb trail has an unnamed link — the error can never reach production.
We also see in our own open dataset how often the machine-reader layer is skipped. The Measurement Observatory measured 160 domains from Turkey and the European Union; of the 131 sites responding live, 47 had structured data and 13 had hreflang tags linking language versions. These sites were selected from a popularity list, so they are large, well-known sites — even among them the machine layer is not the exception, but it is not the rule either.
Every section is written twice: once for the human who reads it, once for the machine that reads it.
If you want to see these answers on your own site, corporate web design delivers design, code and technical SEO from one hand, on a measurable budget.