Business website — which kind makes sense for which company

A conversation about a business website usually starts with "what does it cost". That is the second question. The first is "which kind" — and until it is answered, a quote covers something nobody has defined yet.

This article answers the first one. It contains no amounts; those live in the cost section and are maintained there. It contains what has to be settled before a figure means anything at all.

Four types — they differ by job, not page count

Page count is a consequence of the choice, not a criterion for it. Below is the split we use ourselves when quoting projects.

A brochure site has to lead to one contact. The sale happens in conversation and the site confirms you exist and gives a number. We cover it separately, together with the five things that have to be on it: brochure website.

A business website has to answer repeating questions. Several services that people ask about separately, each with its own address and its own answer. This is the type the rest of this article is about — and the usual choice for service firms.

A larger content site has to carry several content types at once: the offer, the work, knowledge, recruitment, sometimes several markets. Content keeps arriving and somebody has to add it without a developer.

A portal with accounts has to accept something and remember it. The user logs in and comes back for their data. This is a system, not a website — a different cost category, a different timeline and a different maintenance burden. One question settles it: does the user have an account?

Two of these four shade into one another and the boundary is worth marking, because it is where most disagreements about a quote start. A business website becomes a larger content site not when it passes some page count, but when somebody has to add content regularly. If nobody in the company will be publishing anything after launch, you do not have a content site — you have a business website with an editorial system nobody will open, and you paid for the system.

Worth noticing that these four are not stages of development you pass through. A service firm can stay with a brochure site for ten years and be right, not negligent. The reverse: a start-up needing a customer panel from day one does not "grow into" a portal — it starts as one. Treating this list as a ladder produces two expensive mistakes at once: adding pages nobody reads, and deferring a system that was needed from the start.

There is a fifth type not listed here, because it plays by entirely different rules: a shop. Selling goods through the site is a different project, with different fixed costs and different risks, and it is covered in its own section.

Three questions that settle the choice

None is about page count and none is about budget.

Where does the sale happen? 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 the project changes category.

How many questions repeat? Count the ones that come up more than three times a month in your conversations. That is your page list, and it is the only honest criterion for page count we know. The most common error is dividing the site by how the company is organised: a customer does not know your structure and is not looking along it.

Does the site have to accept something? An order, a booking, a service request — not just a form that lands in an inbox. That is the threshold at which the project changes category.

What actually moves the quote

Four things, and only one of them is irreversible.

Scope of content — how many pages and who writes the text. This is the item that cannot be deferred, because the URL structure, the navigation and what a search engine finds all depend on it. Adding ten pages to a site designed for three ends in a menu nobody understands.

The content management system — whether anybody is to edit it without a developer. Swappable later, at a cost.

The level of design — template, custom design, or custom design with motion. Swappable later.

Interactive features — accounts, bookings, search, integrations. Swappable later, and usually better deferred until something actually needs them.

The practical consequence when the budget will not stretch: cut design and features, not scope. That is the opposite of what usually gets cut, because scope is the invisible one — a shorter page list does not look like a compromise in the way a plainer design does, and that is exactly why it gets chosen.

There is a second thing scope decides that rarely appears in quotes: who can answer a customer without you. Each page that exists is an answer somebody in your firm can send as a link instead of retyping. Ten such pages change how the week runs; three do not. That is the return on scope, and it is invisible in a design comparison.

What this site is supposed to connect to

Integration is the item most often quoted and least often needed. Before paying for it, check whether there is anything to connect to.

In Poland in 2025, according to Statistics Poland (GUS), the report "Information society in Poland in 2025", 40.5% of enterprises used ERP software and 25.1% used CRM. Read the CRM figure again: in three companies out of four there is no customer system for a website to talk to.

That is not an argument against integration — it is an argument for checking first. If you have no CRM, integration is not a line in a quote, it is a decision to postpone, and a site that emails enquiries to an inbox is the correct solution until the volume of enquiries outgrows it.

That makes the integration line in a quote worth reading closely: if you are small and your quote contains one, the question to ask is which system, on which side, and who maintains the connection when either end updates.

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

Six things common to all four types. None is an extra, and each gets quoted as an extra.

Its own address, not a contractor's subdomain. Working on a phone, tested on a real phone on mobile data rather than in a desktop preview. Contact visible without hunting, on every page. An encrypted connection, which browsers now enforce rather than reward. The ability to edit content yourself — prices, hours, a phone number. And a backup somebody has actually restored, which is a different claim from "we have backups".

What is not on that list: animation, scroll effects, background video. Not because they are bad, but because they are a choice and these six are a condition.

The one most often missing in practice is the fifth. "You can edit it yourself" is agreed at the start and tested by nobody, and six months later every change to a price goes to the contractor by email and gets billed. The test is cheap and takes five minutes: before the project is signed off, ask the person in your company who will actually do it to change a phone number while you watch.

What stays on your side whatever the contractor does

Three things cannot be outsourced, and they are what most often decides the timeline.

Substantive content. Service descriptions, photographs of work, prices or ranges. A contractor can edit it but cannot invent it — and a project waiting for material stalls for weeks. It is also the largest item you control, including on cost.

Decisions. What you sell, to whom, at what price and what you do not do. It sounds obvious and it is the most common cause of delay, because writing a website is where it turns out they were never taken.

One person who approves. Three people approving a layout means three rounds of revisions and a month nobody planned for.

And three things must be registered to you, not to the contractor: the domain, access to wherever the site is hosted, and the analytics accounts. Checking the domain is free — registrant details are public.

What happens to this site after a year

Three things, in this order, and all three are cheaper to plan for than to discover.

Content goes stale before the design does. Prices, staff, the service list. This is the failure that damages credibility, because a visitor cannot tell an out-of-date page from a careless company.

Software versions come due. Whatever the site is built on will stop receiving security updates on a schedule nobody in your company set. That is a maintenance cost, not an investment, and it returns every two or three years.

Somebody will want a change the site was not built for. A new service, a second language, a booking form. Whether that is an afternoon or a rebuild was decided by the scope question at the very beginning.

The second language deserves a warning of its own, because it is the change most often assumed to be easy. Adding one is not a translation job — it doubles the content that has to stay current, needs its own URL structure, and needs somebody who can tell whether the translated version still says what you meant. Firms that add one "for later" usually end up with a half-maintained second site that damages credibility in both languages.

Five mistakes when commissioning

• Asking for a quote before settling the scope. Two quotes for "a business website" can differ by a factor of two with no difference in margin — they are quoting different things.
• Choosing by the final figure. The cheaper quote is usually the one that assumes you will supply the text, the photographs and the decisions.
• Leaving the domain to the contractor. The only irreversible item on the list.
• No redirects when replacing an old site. Pages that change address without one disappear from search quietly, and nothing looks broken.
• Nobody named to maintain it. A site with no owner inside the company is out of date within a year, whatever it cost.

What connects four of those five is that each was a decision deferred rather than made. Scope deferred to the quote stage, ownership deferred to the handover, maintenance deferred to whoever notices first. A website is unusually good at revealing which decisions a business has been postponing, because it will not launch until they are taken.

Who will build it

The choice of contractor is decided less by portfolio than by three answers, and all three can be got in one call.

Who writes the text. If the answer is "you supply it", that is a cost and a timeline on your side and it belongs in your plan, not in theirs. If the answer is "we can write it from an interview", that is a real service and it should appear as a line.

What happens after launch. Updates, backups, and who answers when something breaks at a bad moment. A contractor who has no answer here is quoting for a build, not for a website.

What gets handed over. A list of versions, accesses and dependencies makes the next project cheaper. Its absence means somebody will rediscover your site from scratch in two years, at your expense.

A quote that arrives without any questions being asked is worth treating carefully. Anyone who can price this without knowing how many services you sell and where your sale happens is pricing a template, and the number will move once the work starts.

How to read a quote

Four things to check, none of which is the total.

Who writes the text and where the material comes from. If the answer is "you supply it", that is a cost on your side and a risk to the timeline.

What maintenance covers and what it costs per year. Hosting, domain, updates, backups, and who responds when something breaks.

Whose the domain and the accesses are. Ask it explicitly and get it in writing.

What happens to the addresses if the new site replaces an old one. The absence of a redirect plan is the defect you do not see until traffic has gone.

How to tell after a year whether the choice was right

Three checks, all free.

Count the questions that still repeat. If people still ask things that are on the site, the problem is findability rather than scope. If they ask things that are not on it, the scope was too narrow.

Check who edited it last. If the answer is "the contractor, at launch", the content management system was a line item that bought nothing.

Compare enquiries against the effort of getting them. A site that produces the same enquiries as before, more slowly, is a site whose scope answered a question nobody was asking.

The shortest version of everything above: decide what the site has to accept and how many questions it has to answer, and the rest of the quote follows. Every expensive surprise we see in this work traces back to one of those two being left open until somebody was already building.

What this article deliberately leaves out

• Amounts — cost.
• Whether a site pays for itself at all — the arithmetic.
• The smallest sensible site — brochure website.
• A page for a campaign — landing page.
• Design and usability — design.
• What to build it on — tools.
• Where traffic comes from — marketing and SEO.