Starting a web design business in 2026 — the full costs, the real margins, and how to run it

Most guides to starting a web design business are written by people selling a course. This one is written by the people who built the tool, which means we have an obvious interest — so everything below is a number you can check, and where the number is uncomfortable we have said so.

The short version: the startup cost is under $100, the per-project cost is under $150, and the constraint is not money or capacity. It is finding clients. Everything else in this guide is arithmetic.

Is web design still a business in 2026?

It is worth confronting the case against first, because you will hear it: AI builders can make a website in minutes, so why would anyone pay a person?

Three reasons, and they have held up.

Small businesses do not want to make a website. They want to have one. The gap between "a tool exists" and "the owner of a bakery sits down and uses it well" is the whole business. Someone has to decide the structure, write the words, judge whether it looks credible, and take responsibility when it does not. The tools got better; the willingness to do the work did not appear.

The number of businesses needing a site keeps growing. New business formation in the US has run well above pre-2020 levels for several years — millions of new applications a year. Most of those are small: trades, food, services, single-operator shops. Every one of them eventually needs a web presence, and most of them still start with a Facebook page and a phone number.

A large share of small businesses still have no site at all, or one they are embarrassed by. Survey figures move around — you will see anywhere from a quarter to a third of small businesses reporting no website — but the direction is consistent and the reason given is always the same two things: cost, and not knowing where to start. Both of which are what you are selling a solution to.

What has changed is the shape of the work. Charging $4,000 for a five-page brochure site is a harder sell than it was, because the client knows tools exist. Charging $600 and delivering in four days, twenty times a quarter, is a better business than it has ever been — because your cost per site has collapsed while the client's need has not.

That is the whole strategic point of this guide: the money is in volume and speed, not in bespoke.

If you want the fuller argument about where AI builders help and where they hurt, we wrote that separately: AI website builders — the honest pros and cons. And on why page builders quietly cost more than they save, there is a piece on that too.

---

Part 1 — What it costs to start

Here is the entire startup cost. There is no hidden tier.

Item	Cost	Notes
Your own domain	$10–15/year	Namecheap, Porkbun, Cloudflare Registrar
Shared hosting (first year)	$30–60/year	See below — this is the one that matters
Business registration	$0–300	Varies enormously by country and structure
GnK Studio, first project	$0	Directions and homepage are free before you pay
Realistic total to be open for business	$40–75	Excluding registration

That is not a rhetorical trick. You can be genuinely operational — your own site, your own domain, hosting ready for a client — for the price of a dinner.

What you do not need: an office, a logo designer, a business coach, a CRM, a project management tool, paid ads, or a course. All of those are things people sell to aspiring freelancers. None are required to deliver the first ten projects.

What you do need that is not money: a way to find clients. That is the actual constraint, and Part 5 deals with it honestly.

---

Part 2 — Hosting: buy shared, and here is exactly why

This is the decision that most affects your margins, and most people get it wrong by buying too much.

Why shared hosting is the right answer

A brochure site for a plumber, a restaurant, a dentist, a landscaper — this is not a demanding workload. It is a handful of pages, a contact form, some images, and traffic that might reach a few thousand visits a month on a good month. Many local business sites see less than that.

Shared hosting exists precisely for this. You are buying a slice of a server alongside other small sites, and for content like this you will never notice the neighbours. Paying for a VPS to serve a five-page restaurant site is like renting a van to carry a laptop.

The honest limitation: shared hosting struggles when a site is genuinely busy or genuinely heavy — a store doing real transaction volume, a site with a huge media library, an application with logged-in users. When a client grows into that, you move them and charge for the migration. That is a good problem and a billable event.

Three providers worth considering

We are naming these because vagueness is useless when you are trying to start. Check current pricing yourself; introductory rates and renewal rates differ, sometimes sharply.

Hostinger — consistently among the cheapest for multi-site plans, with a control panel that is friendlier than cPanel for people who do not live in servers. Their business-tier plans typically allow dozens of sites. The catch, and it is industry-wide: the headline price is an introductory rate on a multi-year term, and renewal is meaningfully higher. Budget for the renewal, not the promotion.

Namecheap — better known for domains, but their shared hosting is inexpensive and their multi-site plans are straightforward. Convenient if you want domains and hosting in one dashboard, which reduces the number of places you can lose a login.

Cloudflare Pages — the outlier, and worth understanding. For static sites it is free, permanently, including HTTPS and global delivery. If you deliver static sites, your hosting cost for those clients is zero. It does not host WordPress, so it is not a general answer, but for the static line it is unbeatable.

The ten-sites-per-account rule

Here is the operating rule that makes the arithmetic work.

Put up to about ten client sites on one shared account, then buy another.

Why ten, and not "as many as the plan allows":

Performance. Shared plans advertise unlimited or high site counts, but you are still sharing CPU and memory. Ten small brochure sites is comfortable. Thirty is how you get slow sites and a client asking why their page takes six seconds.

Blast radius. One account means one point of failure. If that account has a problem — a suspension, a billing lapse, a compromised site — everything on it goes down together. Ten clients calling you at once is survivable. Forty is a business-ending afternoon.

Clean separation. When a client leaves, or grows and needs their own hosting, moving one site off an account of ten is simple. Untangling one from forty is not.

So: account one fills to ten, buy account two. At $50 a year per account, ten clients cost you $5 each per year in hosting.

The hosting offer that costs you nothing

This is the part worth reading twice.

Offer the client: first year of domain and hosting included, then a small annual fee — call it $95 a year — to keep it live, backed up and updated.

Run the numbers from your side, per client per year:

• Domain: ~$12
• Hosting share (1/10th of a $50 account): ~$5
• Your cost: ~$17
• You charge: $95
• Margin: ~$78 per client, per year, recurring

With 10 clients that is roughly $780 a year of recurring revenue against $170 of cost. With 40 clients, across four hosting accounts, about $3,100 a year for maybe an afternoon a month of actual work.

Two things make this honest rather than a trick:

The client genuinely gets value. They get a domain they own, hosting that works, and someone who answers when something breaks. $95 a year is less than most of them pay for a single month of a website builder subscription.

Say what the renewal covers, in writing. Hosting, domain renewal, and keeping the thing online. If you are not doing content updates for that price, say so. The fastest way to poison a recurring fee is to leave its scope vague and then argue about it in month seven.

---

Part 3 — The per-project arithmetic

Now the delivery side.

Your cost per site

Using GnK Studio, a typical five-page WordPress project:

Line	Cost
Base project, up to 4 pages	$99
One extra page	$15
Standard control (client can edit logo, tagline, socials)	$25
White-label delivery (your studio as the author)	$19
Total per project	$158

Buy your balance in advance and it drops. Top up $500 and you get $600 of balance — an effective 16.7% discount — so that $158 project costs you about $132 of real money. Top up $1,000 for $1,250 and it is closer to $126.

Add the first-year hosting and domain you are including (~$17) and your all-in cost per client is roughly $150.

What to charge

Market rates for a small business brochure site from a freelancer run enormously wide — $500 to $5,000 depending on your market, your positioning and the client's size. The volume strategy says: sit deliberately at the low-to-middle end and win on speed and certainty.

You charge	Your cost	Profit per project
$400	~$150	$250
$600	~$150	$450
$800	~$150	$650
$1,200	~$150	$1,050

At $600 — which is a genuinely easy price for a local business that has been quoted $3,000 elsewhere — you keep about $450 per project.

Monthly income at different volumes

Projects/month	At $400	At $600	At $800
2	$500	$900	$1,300
4	$1,000	$1,800	$2,600
8	$2,000	$3,600	$5,200
12	$3,000	$5,400	$7,800

Add the recurring hosting fees on top, which compound as your client base grows and require almost no additional work.

Is 8–12 projects a month realistic? On delivery capacity, comfortably — the studio work per project is a brief conversation and a few approval rounds, not a week of production. On sales, that is the hard part, and anyone who tells you otherwise is selling something. Most people building this take months to get to steady volume. The point is that your delivery capacity is not the ceiling, which is unusual and is the entire advantage.

---

Part 4 — How to actually run it

The workflow that keeps projects short

1. Qualify in ten minutes. What business, how many pages, do they have photos and text, when do they need it. If they cannot answer what their business does in two sentences, the project will be painful — price accordingly or decline.

2. Take a deposit. Half up front is standard and filters out the people who were never going to pay. You are risking about $150 of cost per project; do not risk it on a stranger's good intentions.

3. Run the brief with the client, not for them. Have the studio conversation on a call. They answer the questions, they see the directions, they pick. This does two things: it takes the design opinion off your shoulders, and it makes them a participant in a decision they will later have to live with.

4. Send a preview link, not a screenshot. The client gets a read-only URL, browses the real mockups on their phone, and approves. No account, no attachments, no "can you send it again". This is where projects usually stall, and removing the friction is worth more than it sounds.

5. Batch revisions. Collect all changes into one round rather than fixing things one message at a time. Two rounds per page are included; make them count.

6. Deliver, install, invoice the balance. Then set the renewal reminder for eleven months out.

Where the time actually goes

Most of your hours will not be design. They will be: answering "can you also…" questions, chasing content the client promised, and finding the next client. Budget accordingly, and notice that the first two shrink dramatically when you set expectations in writing at the start.

Finding the first ten clients

Nobody can hand you this. But the pattern that works for local service businesses is boringly consistent:

Start with businesses you already have a relationship with. Your barber, your dentist, the restaurant you go to weekly. You are not cold-calling; you are offering something to someone who already knows your face.

Look for businesses with a bad site, not no site. A business with no site has not decided it needs one. A business with a broken, ugly, or mobile-hostile site has already decided — they just have not found someone affordable.

Show, do not pitch. Build the site first for one or two of them, free or nearly free, and show them the preview link. "Here is what yours could look like" outperforms any sales conversation.

Ask every satisfied client for one introduction. Not a testimonial, an introduction. This is how small local businesses actually find suppliers.

---

Part 5 — The honest risks

We would rather you go in clear-eyed.

Sales is the bottleneck, permanently. Delivery is solved. Finding people who want to pay is not, and it does not become easy. If you dislike selling, this business will feel like grinding.

Cheap clients can be expensive. The client paying $400 sometimes wants more attention than the one paying $2,000. Scope in writing, and enforce it kindly but actually.

Renewal fees need administration. Forty clients paying annually means forty invoices, some chasing, and some churn. It is good revenue but it is not free revenue.

Your suppliers are dependencies. Hosting companies change prices. Platforms change rules. Build the habit of keeping client files in your own storage so that any single supplier failing is an inconvenience rather than a catastrophe.

Do not oversell what you deliver. Selling a static site to a client who wants to blog weekly, or a brochure site to someone who needs a booking system, ends in a refund and a bad reputation. Matching the right line to the real need is most of the skill.

---

The summary

• To start: about $50, plus a domain.
• Per project: about $150 all in, less if you prepay your balance.
• Charge: $400–$1,200 depending on your market.
• Recurring: ~$78 per client per year, for very little work.
• The constraint: finding clients — never delivery.

Ten clients at $600, plus their annual renewals, is roughly $5,300 in the first year and $780 a year afterwards from doing nothing further. Forty clients over two years, at the same rates, is a real small business.

None of that requires a team, an office, or an investor. It requires being able to find small businesses that need a website, and being reliable enough that they tell someone else about you.

Ready to run one? Start a project — the three design directions and your client's homepage cost nothing to see, so your first pitch can be a finished mockup rather than a promise.