Ask five studios to build the same site and the quotes will land an order of magnitude apart. That spread is not fraud and it is not incompetence. It is the absence of a shared definition: “a website” describes an outcome about as precisely as “a vehicle” describes a purchase.
So here is the breakdown that usually stays inside the studio.
Why one brief becomes five different prices
Before the numbers, the reason they diverge. Every studio reading your brief silently fills in the blanks, and they fill them in differently.
- Who writes the content. One studio assumed you will deliver finished text. Another priced a copywriter. That single assumption can be a fifth of the total.
- How much design is new. A template adapted to your brand and an original design are both legitimate answers to “make it look good”, and they are not close in price.
- How much testing is included. One quote covers a pass on real devices and a fix cycle. Another covers “it works on my laptop”.
- What happens to your old data. Nine hundred products and four years of orders do not migrate themselves, and whoever ignored that line is cheaper on paper.
- Who carries the risk of the unknown. A fixed price includes a buffer for surprises. An hourly rate moves that risk to you and looks lower at the start.
None of these makes a studio dishonest. It makes the totals incomparable until you force the assumptions into the open, which is the single most valuable hour you will spend on procurement.
Where the money actually goes
On a custom project that goes well, the budget splits roughly like this:
- Deciding what to build — 10 to 15%. Interviews, digging out the real business task, saying no to features. The phase clients most want to skip, and the one that most often halves the cost of everything after it.
- Design — 20 to 25%. Not decoration: the layouts, states and rules that the build then follows without improvising.
- Build — 35 to 40%. The part everyone pictures when they imagine paying for a website.
- Content and data — around 10%. Getting real text, real images and old data into the new thing. Almost always underestimated, usually by the client, because it is the part nobody enjoys.
- Testing, launch and handover — around 10%. Including the boring items that decide whether you sleep after launch: backups, monitoring, analytics, access.
Now look at what is invisible in that list. The first item and the last two — roughly a third of the budget — produce nothing you can screenshot. That is precisely where cheap quotes get cheap. Nobody advertises “we removed the testing”; they simply quote a number that cannot contain it.
What you are buying, by shape
Pages are a poor unit. Twenty pages of the same template cost less than one page with a configurator on it. The useful bands are defined by what the thing has to do:
- A presence. A good template, your content, sensible setup. You are buying speed to market, not distinctiveness.
- A custom small site. Original design, a handful of unique layouts, a content system your team can actually operate. You are buying distinctiveness and control.
- A site with a job. It captures leads, talks to your CRM or booking system, and is measured on conversion. You are buying a machine that is supposed to pay for itself.
- A store. Catalogue, checkout, payments, tax, shipping, returns, email flows. The common mistake is believing the storefront is the project; the machinery behind it is.
- A product. Accounts, subscriptions, permissions, an admin nobody dreads opening. This is software, it is priced like software, and it never really finishes.
Which band you are in moves the price far more than page count, and any studio should be able to place you during a first conversation. If they cannot, they are pricing pages.
The three line items that are usually padding
- Price per page. Templates are cheap to repeat. Charging per page bills you for copy and paste. Ask which pages are genuinely unique, and pay for those.
- The “SEO package”. Frequently a plugin, a keyword list and a monthly report. The SEO that works is structure, speed, crawlability and content — mostly inside the build, not sold alongside it.
- Unlimited revisions. Priced into the base with a buffer for the worst client the studio ever had. Decisive clients subsidise indecisive ones.
The four things that genuinely move the price
- How many people have to approve. Two decision makers is a project. Six is a negotiation, and negotiations are billed by the hour whether or not anybody says so out loud.
- Whether your content exists. Real text, real photography, ready to use. This is the single most common reason projects cost more and take longer than the quote said.
- Integrations with systems you do not control. Every external system is a small project with unknown edges: an old interface, missing documentation, a vendor who replies next week.
- Whether anyone has said no yet. A feature list nobody has pruned is the most expensive document in the process, and pruning it is the cheapest work anyone will do for you.
What to cut when the budget will not move
In this order, and we will say it out loud in a first call: page count first, then custom illustration and photography, then animation and polish, then flexibility in the content system — fewer editable areas, more decisions made once and fixed.
What we will argue about rather than cut: testing, analytics and the handover. Cutting testing does not remove the defects, it moves them to your customers, who report them by leaving. Cutting analytics means you will never learn whether any of this worked, which turns the whole purchase into an act of faith. Cutting the handover means paying somebody forever for access to your own website.
When we tell people not to hire us
If a budget covers the build and nothing else — no content, no photography, nothing left to bring anybody to the site — then a good template plus that spending beats a custom site nobody visits. We have given that advice more than once. It costs us the project and saves the client a year. A beautiful site with no traffic and no product photos is an expensive way to discover what your actual problem was.
The same logic runs the other way. If the site is already earning and the complaint is “it feels dated”, the money often belongs in three targeted fixes rather than a rebuild — we made the whole argument for not rebuilding elsewhere, and it costs us bigger projects than this one does.
What a studio should ask before it quotes anything
A price that arrives without questions is a guess wearing a suit. Before quoting, anybody serious needs to know what the site is supposed to change in your business, who signs off, what content exists today, which systems it has to talk to, and what your real deadline is driven by — a season, a trade show, a contract, or a preference.
If a quote arrives without those five answers, one of two things is true: the studio has priced a generic project and will discover yours later, or it has quietly padded the number to survive whatever it finds. Both are expensive, and the second one is expensive even when nothing goes wrong.
The number that actually decides this
The right budget is not the market average. It is your own arithmetic: what should this site add over the next twelve to eighteen months — orders, qualified leads, hours removed from someone's week — and what is that worth in money? A site expected to add fifty thousand in margin justifies a different number than one expected to add five, and neither number appears in anybody's price list.
Once you have it, ask each studio two things: which band they think you are in, and what their day rate is. Two studios with the same day rate quoting very different totals are describing different amounts of work — and now you can ask exactly which work one of them left out.