What actually drives the cost of a custom website

You want a number. I understand — you're trying to work out whether this is a two-thousand conversation or a twenty-thousand one before you invest time in it.

The reason you won't find a credible number in an article is that "a custom website" describes both a five-page site for a physiotherapist and a multilingual store with an ERP integration. Any figure specific enough to be useful is wrong for most readers. What is transferable is the structure underneath the number: what makes a quote go up, what makes it go down, and how to tell whether two very different proposals are even describing the same project.

Why nobody can quote you from a sentence

"I need a website for my business" isn't a brief, it's a category. Two clients who send me that exact sentence can turn out to need work that differs by a factor of ten:

• One needs five pages, a contact form and their existing text loaded.
• The other needs eighteen pages in two languages, a booking system connected to their calendar, a member area, and photography they don't yet have.

Same sentence. Any developer who gives you a firm price before knowing which one you are is either padding heavily to cover the risk, or planning to discover the rest of the scope after you've signed. Neither is good for you.

The seven variables that move a quote

When I price a project, these are the things I'm actually measuring. They're worth knowing because you control most of them.

1. Page count — but really, page types

Twelve service pages built from one template is barely more work than one. Six pages that each need a different layout is a much bigger job. When you're estimating your own project, count distinct designs, not URLs.

2. Where the design comes from

Three different situations, three different prices: you have an approved Figma file (cheapest — the decisions are made); the design is created as part of the project (middle); or the design is created collaboratively with several rounds of stakeholder feedback (most expensive, and the one that most often overruns).

3. Functionality beyond pages and forms

This is the biggest single lever. A brochure site is a known quantity. Ecommerce, bookings, memberships, user accounts, quote calculators, integrations with a CRM or an accounting system — each of these is a project inside the project, with its own testing and its own edge cases.

4. Content: who writes it, who shoots it

If you deliver finished text and usable images, that's work nobody has to bill you for. If someone has to write the copy, structure it, source or shoot photography, and edit it — that's a real and separate cost, often a surprising share of the total.

5. Languages

A second language is not "the same site again." It's translation, a URL structure, hreflang tags, duplicated content maintenance, and a layout that survives German nouns and Spanish sentences that run 20% longer than the English. It's the variable people most often forget to mention up front.

6. Migration and existing baggage

A new site on a blank domain is straightforward. Replacing a site with years of content and rankings means an audit, a URL map, redirects, and preserving what already earns you traffic. That's careful, unglamorous work and it takes real time.

7. Timeline

A normal schedule costs what it costs. A compressed one costs more, because it means declining other work or working weekends. If your deadline is real, say so at the start — it's much cheaper to plan for than to rescue.

Why two developers price the same brief differently

You send an identical brief to three people and get three numbers far apart. That's not necessarily anyone being dishonest. The gap usually comes from:

• Different assumptions about what's included. One priced content loading; another assumed you'd do it. One included analytics setup and training; another didn't.
• Different approaches. A template-based build and a bespoke one are legitimately different products at legitimately different prices.
• Different revision policies. A quote with two rounds included and a quote with "unlimited revisions" are not comparable — the second has the risk priced in, or will be renegotiated.
• Different amounts of risk absorbed. Someone who has read your brief carefully and spotted the awkward part has priced it. Someone who skimmed hasn't — yet.
• Different costs of living and demand. A real factor across markets, and not a proxy for quality in either direction.

Costs that aren't in the build price

The build is a one-off. These continue, and they belong in your decision:

• Hosting, which varies enormously between shared hosting and a managed platform.
• Domain renewal, small but perpetual.
• Licences for premium themes, plugins or apps, most of which renew annually.
• Platform and app subscriptions on hosted platforms — this is the one that grows quietly.
• Maintenance: updates, backups, monitoring. Either you pay someone or you spend your own time.
• Changes: the new service page, the seasonal banner, the form field. Budget something for the year, because there's always something.

A quote that ignores all of this isn't cheaper. It's just less complete.

How to compare three proposals fairly

Put them side by side and check they're describing the same thing:

1. Same page count and page types?
2. Who loads the content — you or them?
3. How many revision rounds, and what counts as one?
4. Is the design custom or template-based?
5. Is training or a handover video included?
6. Is there a warranty period for bugs after launch, and how long?
7. Who owns the code and design when it's paid for?
8. Whose name is on the hosting and domain accounts?

Answer those eight for each proposal and the numbers usually stop looking so different. Where a gap survives the exercise, you're now comparing approaches rather than guessing.

What the cheapest quote is usually leaving out

Cheap isn't automatically bad — an experienced developer with a template they know well can deliver quickly and price accordingly. But a quote far below the others is usually missing something specific:

• A licensed template with your logo dropped in, described as "custom."
• Content loading excluded, discovered when you're asked to fill fifteen pages yourself.
• No mobile-specific work beyond whatever the theme does by default.
• No SEO structure: no meta descriptions, no heading hierarchy, no redirects from the old site.
• No testing beyond one browser on one screen size.
• No handover, so every future change comes back to them at an hourly rate.
• Hosting on their account, which is where "cheap build" becomes "expensive to leave."

The way to tell is to compare the itemised scope, not the total. If a proposal has no itemised scope, that's the finding.

How to spend less without getting less

Most of the levers are on your side of the table:

Deliver your content first, in one batch. Finished text and usable images remove the single largest source of delay and rework on any project.

Cut page types, not pages. Twelve service pages on one template is cheap. Four pages with four bespoke layouts is not.

Launch smaller, then grow. A tight, excellent eight-page site that goes live in a month beats a twenty-page site that's been "nearly ready" since spring. Phase two can be funded by phase one working.

Nominate one decision-maker. Revision rounds are where budgets die, and they multiply when four people give contradictory feedback.

Be honest about the deadline. Inventing urgency you don't have costs you money for no benefit.

Don't buy features on speculation. The member area you might want next year will be cheaper to build next year, when you know whether you want it.

FAQ

Why won't developers publish their prices?

Some do, for standardised packages where the scope really is fixed. For custom work, a published price is either high enough to cover the worst-case project — which overcharges everyone with a simple one — or low enough to attract enquiries it can't actually deliver. A quote after a short conversation is more accurate for you, not less.

Is a freelancer cheaper than an agency?

Usually, because you're not funding project managers, account handlers and office overhead. The tradeoff is capacity: one person has a limit on parallel work and no redundancy if they're ill. For most small and medium projects the freelancer maths works out; for large multi-discipline programmes it often doesn't.

Should I pay hourly or a fixed price?

Fixed price for a defined project, because it puts the estimation risk on the developer and gives you a number you can plan around. Hourly for open-ended work — ongoing improvements, maintenance, "we'll see how it goes" — where a fixed price would just be a guess with padding on top.

What deposit is normal?

A deposit before work starts is standard practice and protects both sides; the rest tied to milestones such as design approval and launch. What isn't standard is the full amount up front before anything exists.

The short version

The price of a custom website is driven by page types, where the design comes from, functionality, content, languages, migration and timeline. Get three quotes on the same written brief, compare the itemised scope rather than the totals, and include the recurring costs in the decision. Then pick the proposal that shows the clearest understanding of your project — not the smallest number.