Ask what a website costs and you will be quoted anywhere from nothing to six figures. It sounds like the industry is making numbers up. It is not. The word “website” is carrying at least four different projects, and nobody in the conversation says which one they mean.
Part of the problem is that the thing itself stopped being what the word describes. A website was something you had in 2004 — an address, some pages, a contact form, done. That is not what a business has now.
What a business has now is a digital presence. A website is one part of it, and for a lot of businesses it is not even the first part.
Some businesses are only a social account
And it works. There are operators doing real revenue whose entire presence is a Facebook page, or TikTok, or Instagram. No site at all. They have poured years into it, they post constantly, and the audience is genuinely theirs in every way that matters day to day.
We are not going to pretend that is a mistake, because for a lot of them it plainly is not. If your customers find you where you already are, and the phone rings, that is a working digital presence.
There is exactly one problem with it, and it is not a small one.
The catch
You are building on land you do not own.
The algorithm changes, the reach halves, the account gets flagged by a system with no phone number, and every relationship you built runs through a door somebody else controls.
That is the actual argument for owning something — not that social is beneath you, and not that a website is more professional. It is that a rented presence can be taken away by a policy update you will never be consulted about.
So the useful question is not “do I need a website.” It is: how much of my presence do I want to own, and what does each step of owning it cost?
The presence ladder
Four rungs. Each is a legitimate place to stand, and each answers a different question. Most cost disputes are two people standing on different rungs using the same word.
Borrowed
Social only · $0 plus your time
You exist where the audience already is. Costs nothing but effort, and the effort is real. Right answer for a new operator, a personality-led business, or anyone testing whether there is demand at all. You own none of it.
Owned address
Brochure site · DIY, or $3,000–$10,000 built properly
Who you are, what you do, how to reach you, and a place that is yours. It presents. Content goes out, enquiries come back.
Be honest with yourself here: this rung can be DIY. Squarespace or Wix, a weekend, and a small business has a perfectly respectable presence. We will tell you that rather than sell you something. What you buy by paying for it is positioning and copy that actually sound like the business, and a build that will not quietly fall over.
Working presence
$10,000–$20,000 · our most common engagement
The site stops presenting and starts doing: booking, commerce, a content operation, search that brings people in without paying for every click, and measurement that shows what happened. Historically our work here has landed between $12,000 and $17,000, and scope moves it inside the band from there.
Most of that budget is not pixels. It is repositioning and copywriting — deciding what the business actually says. That is usually where the value of a rebuild comes from, and it is the part people are most surprised to be paying for.
Interactive presence
$25,000+ · from interactive microsite to full ecosystem
Accounts. Members doing things to each other’s data. Dashboards over live state. A community rather than an audience. This rung has its own range, and quality is what moves you along it — from an interactive microsite at the bottom to a full-fledged ecosystem at the top: a digital platform running on AI and integrated systems.
This is where the word “website” stops being useful, and where budgets detonate.
Where presence becomes platform
Rung four is not a more expensive rung three. It is a different discipline. Different testing, different security exposure, different failure modes, and an ongoing obligation a brochure never carries. A brochure that breaks is embarrassing. An application that breaks locks people out of their own data.
| Still a website | Now an application |
|---|---|
| Pages, blog, service detail | User accounts and permissions |
| Forms and enquiries | Members interacting with each other |
| A catalogue and checkout | Dashboards over live data |
| Booking a time | Workflow, state, and approvals |
| Content managed by staff | Content created by members |
We do not think a website should ever cost more than $50,000. Past that number, what is being described is app development, and it deserves to be scoped, priced and staffed as app development. We build those too. We just refuse to call one the other.
This is the confusion doing the most damage right now, and AI has made it worse rather than better. People have been told software is easy to build now, so the ask arrives larger — a community, a portal, a marketplace — still wearing the word “website,” still carrying a website’s budget expectation.
Three firms quote the same brief. All three are honest. None of them are comparable.
What the word costs you
One quotes $6,000, having heard rung two. One quotes $40,000, having heard the word “portal.” One quotes $90,000, having understood the whole thing. The buyer concludes the industry is a racket, and picks the cheapest — which is the quote from the firm that understood the least.
The fix is not a cheaper number. It is naming the rung out loud before anybody prices anything.
The bill nobody quotes
Whichever rung you land on, the build price is not the price. A quote covers getting live; it does not cover staying live, and those arrive on different schedules.
| Line | When it hits | What people assume |
|---|---|---|
| Hosting | Renewal, year two | The intro rate is the rate |
| Theme & plugin licenses | Annually, each | One-time purchase |
| Maintenance & security | Monthly, forever | Included in the build |
| Marketing & SEO tools | Monthly, each | Optional |
| Payment processing | Every single sale | Rarely modelled at all |
Processing is the one that surprises people, because it scales with success rather than sitting still. Stripe’s published US rate is 2.9% plus 30¢ per successful card charge. A store doing $50,000 a month hands over roughly $1,450 a month before a single license renews — north of $17,000 a year.
So compare three-year total cost of ownership, never launch quotes. On a conventional stack the build is frequently a third to a half of what the thing actually costs by year three.
What AI actually changed
It moved the floor. It did not move the ceiling.
Work that could not have been done responsibly for $5,000 five years ago can start at $3,000 now. But complexity was never about typing speed — it was about decisions, integrations, and the review that keeps it correct, and none of those got cheaper.
Which is the whole distinction we would want a client to hold onto. There is a flood of sites now that are a machine’s output shipped as-is, and they are cheap for a reason: nobody is accountable for them.
It is not an AI website. It is a website built with AI.
The difference is who answers when it breaks
Which rung are you on?
Tell us what you want it to do. We will tell you which rung that is, and say so plainly if the answer is that you do not need us yet.
Basis & limits
- This is a framework, not an audit. The ladder is how we think about scope, argued from building at every rung. It is not a measured finding and there is no dataset behind it.
- Our bands are real but not a quote. They describe where our work has landed and where we expect it to land. A specific project can sit outside them for reasons that only appear once scope is on the table.
- The $50,000 line is a judgement. Someone could reasonably put it at $40,000 or $70,000. What we are confident about is that a line exists, and that pretending otherwise is what produces quotes that cannot be compared.
- Rung one is not a straw man. Social-only presence genuinely works for a lot of businesses and we are not claiming otherwise. The ownership risk is real, but it is a risk, not a certainty.
- Third-party figures are cited, not ours. The card processing rate is Stripe’s published US pricing at time of writing and should be re-checked. The build-price-versus-year-two framing is well covered elsewhere; Codeable’s WordPress cost breakdown is a good treatment and informed that section.
- Recurring cost on our own platform is not stated here yet. We will not publish a monthly figure before it is settled, and a cost article that quietly omits its own recurring line would be doing the exact thing this piece objects to.