Website design cost — why two quotes for the same site differ sixfold

Website design cost has one property that sets it apart from most things a company buys: two quotes for the same thing can differ sixfold and both can be honest.

You can watch it happen on a single set of offers. The same brochure site — five pages, a form and a gallery — is quoted at PLN 3,000 by one supplier and at PLN 18,000 by another, a sixfold difference. The two prices describe two different processes, and neither offer says so.

Below are the six factors that produce that spread, strongest first. One note on the numbers first: all the figures here come from our own study of 112 offers on the Polish market (May 2026), taken from publicly available price lists, with no company names. The full breakdown is in our report. We state the number of observations behind each row, so you can see how far each figure can carry.

Who is quoting — the strongest single factor

Stronger than the technology and stronger than the location. The same scope priced by four kinds of supplier spreads further than by anything else.

Our study splits the offers by kind of supplier, and the ladder looks like this:

The builder row rests on one observation, so read it as an order of magnitude only. The software house median is more than seven times the agency median — that gap, not the entry points, is the finding.

This is not one service under four names. A software house prices a project with integrations, its own business logic and an architecture built to be extended; the freelancer prices a site. Between them sits the bulk of the market — interactive agencies, 61 of the 112 offers — and that is where most quotes come from.

The practical conclusion: if two quotes differ fivefold, first check whether they come from two different rows. If they do, the difference sits in scope, not in margin.

What PLN 3,000 buys and what PLN 18,000 buys

This is the heart of the spread and the thing no quote spells out. Both figures buy "a website". They buy completely different processes.

At PLN 3,000 one person usually does everything: picks a theme, adapts colours and content, deploys. No separate design stage, no testing across devices beyond a quick look, nobody who owns the question of whether the site brings in enquiries. There is a working website — and for many companies that is exactly what they need.

At PLN 18,000 a team works on it, and that is the whole difference. Someone runs the project, someone designs a layout for a specific goal, someone builds it, someone tests it, someone owns performance. A design is made from scratch rather than a template adapted, and after launch there is somebody to call.

So the difference is not that the second site is six times prettier. It is how many people and how many stages stand between the brief and publication — and who carries the risk when something goes wrong.

The question to ask when comparing quotes is therefore not "why so expensive", but: how many people work on this project and which stages are in the price. That one answer explains most of the spread.

Scope: three orders of magnitude, not three brackets

The second factor is obvious until you see the scale. The type of site is a difference of category, not of degree — like comparing a garage with a tower block.

In our study the median brochure site costs PLN 2,500, a business site PLN 5,900, and an extended site with integrations PLN 15,000. Above that begin the portals and enterprise builds, where the budget conversation looks different and cannot be reduced to a single median.

Page count grows non-linearly. The fifth page costs little, because it repeats a pattern already designed. The thirtieth costs more, because at that scale you acquire information architecture, categories, search and the question of how anyone finds anything. Where the offers price the marginal page, above fifteen pages each further page costs PLN 200–500: small per page, and not small by page thirty.

Template or a design from scratch

The third factor is the easiest to miss, because in a quote both versions read identically: "website design".

This is the one factor you can deliberately choose downwards without losing anything technical. The difference is design work, not code. In our study the same functionality costs PLN 3,500 on a ready template and PLN 9,000 designed from scratch — a factor of 2.5. A template means the layout exists and the work is fitting it to your content and brand; a design from scratch means the layout is built around your customer's path.

Which makes sense when that path is unusual, and does not when it looks like everyone else's. Honestly, here is the boundary: for a business site with a contact form a template is usually enough. For a site meant to carry someone through several steps to one conversion, bending a template can cost more than designing from nothing.

Technology — a decision made for you that you pay for

The fourth factor is normally settled on the supplier's side, and it lands on your invoice.

In our study technology is a factor of 3.2 at the same scope. The build price by system:

Treat the Next.js row as an order of magnitude, not a negotiating anchor: it rests on four offers.

What is defensible is the direction. A freelancer proposes WordPress; a software house proposes something written from scratch — and both can be right, because they answer different questions. WordPress is cheaper to deploy and sits in an ecosystem where nearly everything already exists. A custom build costs more at the start and pays back when the site has to do something ready-made components do not. Above both sits a third floor — Drupal, Laravel, headless — no longer websites in the sense this article uses; they are software and are priced as software.

None of this says one is better. That choice — on criteria, not prices — we settle in the technology section.

Five questions that settle the technology

The stack is often presented as a technical decision and therefore taken without you. In practice five business questions settle it, and the answers are yours, not the supplier's.

1. What is this site meant to do beyond presenting the offer? If the answer is "nothing", the upper floor is spending without cover.
2. Who will change the content, and how often? Daily editorial work by a non-technical person is a different requirement from updating a price list once a quarter — and that, not appearance, decides the system.
3. What does it have to connect to? Every integration pushes the project towards a higher row. Three integrations usually push it out of the row entirely.
4. What traffic do you expect in two years? Not today — in two years. Scaling up is cheap; rewriting from scratch is not.
5. Who maintains it when the engagement ends? A technology with no available suppliers in your city or your budget is expensive regardless of what the build cost.

The rule that orders those five answers: take the lowest floor that holds all your yeses. A floor up can always be bought later; going down cannot.

Where the supplier works

The fifth factor is usually dismissed as unserious, and in Poland the gap is inside the country: the same scope costs PLN 15,000 in Warsaw and PLN 3,700 in Wrocław or Poznań.

The reliable contrast is Warsaw (11 offers) against Wrocław and Poznań (19): the medians differ fourfold. The cause is the cost of doing business and what clients can pay locally — and that difference is real, and still not a difference in what you receive. With remote work — and websites have been built remotely for years — this is the only one of the six factors with no connection at all to what you get.

The practical consequence: when comparing quotes, know where they come from. Not to pick the cheapest, but so that you do not read a regional price difference as a difference in scope.

The stage that begins before design

In cheaper quotes this stage does not happen at all, and that is one of the things that explains the spread — because you cannot see it in the description or on the finished site.

It is the work before the first design: establishing what the site is meant to do, for whom, and what the path from arrival to contact looks like. Reviewing a sector's competition is usually a dozen hours or more — the leaders' sites, the queries they answer, how they order their offer, where they put the call to contact. Then information architecture: how many clicks separate arrival from the form, and what must be visible without scrolling.

At the rates on this market the arithmetic explains the missing stage. A dozen-odd hours of competitor analysis comes to roughly PLN 1,000–3,000 — as much as the whole of a budget quote. That is why it is absent: an analysis stage does not fit a budget that barely fits the build.

Is the analysis stage always needed? No. For a one-pager that gives opening hours and a phone number, competitor research changes nothing. For a site meant to produce enquiries it is the only stage that decides whether it does — the rest of the project executes its findings.

Work you cannot see in the quote

The sixth factor is the items that appear as one line, or not at all, and can move the bill by a large fraction.

Integrations. Every one — CRM, booking, stock, payments — is priced as its own piece of work, because the risk sits in the other system's API rather than in your site. Three of them usually move a project into the next row above, and this is the commonest reason two firms quote "the same site" at double the price: one counted the integrations, the other asked about them later.

Content. Text and photographs are the only part you cannot buy along with the site unless ordered separately. Builds sit for weeks waiting for material nobody prepared — and if the supplier writes it, that is a separate line.

SEO at launch. URL structure, metadata, sitemap, technical basics. Done during the build they cost a fraction of what they cost afterwards, because afterwards means changing addresses and adding redirects.

The three share one property: they do not disappear if you do not buy them. They move onto you, or onto later, usually at a higher price.

Why a midpoint is not a price

Every figure here is a median from our study of 112 offers. That decides how slowly it may be read — worth naming before anyone drops it into a budget.

A median is the middle of the market, not a price and not an average: half of the offers are cheaper, half dearer. The range "half of the offers" marks where the middle half sits. Arithmetic means would be higher, because single enterprise builds in the hundreds of thousands pull them up — which is why we give none. For segments, the budget/mid boundary falls at about PLN 3,500 and the mid/premium boundary at about PLN 10,000.

Sample size matters as much as the amount. That is why every row above states its number of observations: the business-site median rests on 66 independent observations, the builder row on one. A figure with few observations shows an order of magnitude, not a negotiating anchor, and we would rather say so than tidy the tables into looking authoritative.

Hence one practical rule: if a quote from one bracket carries a price from another, it is not a bargain — it is a sign that scope means different things to the two of you.

Time — the factor nobody lists, and everybody pays for

A deadline gets treated as an organisational parameter when it is a line item. The same scope delivered in eight weeks and in two is not the same work: compressing the schedule means running stages in parallel that normally run in sequence, and that takes more people at once.

It reaches the bill two ways. The first is obvious: urgent work costs more, because it displaces other projects. The second is less visible and costs more — under a compressed deadline the design stage goes first, because it is the only one skippable without immediate consequences. Those arrive later, as a site that looks good and leads nowhere.

Worth knowing where deadlines actually slip, because it is almost never the coding. Builds stall on content — on text and photographs nobody on the client side prepared, because in the schedule they looked like a detail. That is also the only stage you can speed up yourself, without paying anybody.

If the deadline is fixed, say so in the enquiry, not at signing. A quote with a deadline and one without are two different quotes, and the gap can exceed the gap between a template and a custom design.

When the factors stack

The six do not act separately, and that is the answer to the question at the top — where the difference between PLN 3,000 and PLN 18,000 for the same brochure site comes from.

The cheaper path is one profile, not a set of independent savings: a freelancer, a ready-made template, WordPress, a lower-cost market, no integrations, client-supplied content. Each lowers the price and they occur together, because they describe one way of working.

The dearer path is also one profile: a team, a design from scratch, a custom stack, an expensive market, integrations, content on the supplier's side.

The caveat without which this list would mislead: the factors must not be multiplied. A large-city software house working in a custom stack is not three independent multipliers; it is one supplier profile in which those traits occur together. That is why the gap between the extreme quotes is sixfold rather than fifty-fold, even though each factor on its own can move a price by two or three times.

The conclusion is simpler than it looks: you are not choosing six things separately, you are choosing one way of working. So the question is not "where do we save", but "which profile describes what I need" — and that depends on whether the site has to exist, or has to do something.

How to read two quotes side by side

The whole spread comes down to four questions asked before the amounts are compared, not after.

1. Which row does each quote come from? Freelancer, agency or software house. Without that you are comparing midpoints from two different markets.
2. How many people work on the project and which stages are in the price? Design, testing, performance, care after launch — they are either there or they are not.
3. Template or a design from scratch? If the quote does not say, that is the first question to ask.
4. What about integrations and content? If one quote counts them and the other is silent, they are not two quotes for the same thing.

Only after those four does the difference between the amounts mean anything. A separate decision worth taking in parallel is the billing model — whether you pay for the build once or monthly; we costed both over five years. How to prepare a brief and run the comparison step by step we set out separately — and what starts accruing the day after launch has its own article.