I’ve been meaning to write this one down for about a year. Every time I sat down to do it, the answer changed. The tools got faster, then faster again, and a number I’d have said out loud in January would’ve been wrong by April. So I kept not writing it, which is a very good way to never write something.
But the question doesn’t stop coming. Every founder who emails asks some version of the same two things: what’s this going to cost, and when could it be live? I’ve been quoting those two numbers for twenty years. In 2026 the honest answer is stranger than it used to be. One of them got a lot smaller. The other one didn’t move at all.
The numbers everyone else will give you
If you search this question you get the agency ranges, and they’re not lying, they’re just wide. Teamvoy’s 2026 guide puts a small-business redesign at $15,000 to $40,000, mid-market at $40,000 to $100,000, and enterprise at “regularly exceed $100,000 and can go far higher.” Clique Studios says a 10 to 20 page site takes eight to twelve weeks. Freelancers come in under all of that, and the big brand-led shops come in well over it.
So a startup marketing site, the eight-to-fifteen-page kind with a homepage that has to explain the product in one scroll, lands somewhere between “a used car” and “a new one.” Helpful. Sort of.
Here’s the thing those ranges hide. They’re priced as if a website is one object you buy. It isn’t. It’s five different things wearing one invoice, and they’ve stopped costing the same amounts.
What you’re actually paying for
When I break a redesign down, it’s these five lines, every time:
The deciding. What is this site for, who is it talking to, what’s the one thing it has to make them believe. Which pages exist. Which ones stop existing. This is the part founders think is already done, because they’ve been living inside the company for two years. It’s almost never done. The pitch deck knows what the company is. The website usually doesn’t.
The words. Every headline, every product description, the thing the button says. Not “copy” as a deliverable. The actual sentences that make a stranger nod.
The design. How it looks, how it moves, what it feels like to be on it. The part everyone assumes is the whole price.
The build. Turning the design into a fast, working site that doesn’t fall over. The part that used to be half the price.
The launch and the keeping. Domain, hosting, redirects so old links don’t die, analytics, the first month of small fixes nobody could have predicted.
Now. Twenty years ago the build was the expensive line, and it deserved to be. A site was hand-cut HTML and a CMS you fought with. Ten years ago it was a theme you bent until it broke. In 2026 I design a site and build it in the same week, often the same day, because the production part got fast in a way that’s hard to explain to someone who didn’t live through the slow version. Digital Applied’s 2026 pricing data has AI-assisted development cutting median build times by 22 to 34 percent year over year, and honestly, for a marketing site, that undersells it.
- The deciding 12%30%
- The words 10%22%
- The design 28%26%
- The build 42%12%
- Launch & keeping 8%10%
But notice what they found next: the hours agencies saved didn’t turn into a discount. They got poured back into “design iteration, performance tuning, and analytics instrumentation.” Same bill, better site. Which is fair. It’s also the polite version of something I’d say more bluntly.
That’s the pivot I didn’t expect when I sat down to write this. I thought I was writing a pricing guide. What I’m actually writing is that the cost of a startup redesign in 2026 is the cost of getting a founder to decide what the company is, out loud, in sentences, and the build is the cheap thing that happens after.
What it takes
Same story with time. Clique’s eight-to-twelve weeks is a real number, and their own list of what stretches it is the tell: “client feedback delays,” “content not ready,” “decision-making challenges (multiple stakeholders, conflicting feedback).” Not one of those is a design or engineering task. They’re all waiting.
I’ve watched it from both sides. A six-week plan becomes eleven because the words came in week seven, because the co-founder who wasn’t in the kickoff has notes in week nine, because the legal review for a pricing page nobody mentioned takes two weeks of its own. The building was never the long pole. It just used to be long enough to hide the waiting behind.
Every week of work is on the calendar. Nothing is waiting.
So when a founder asks “how long,” the honest answer has two halves. Four to six weeks when the decisions come easy, when there’s one person who can say yes and the words show up on time. Closer to eight to twelve when they don’t, and that’s nobody’s fault, it’s just where the weeks go. The site is the same site either way. The difference is entirely whether the deciding happens on the calendar or in the gaps. I wrote about the calendar version in Before the Meeting Ends, where a fix went live during the call that surfaced it, and that’s not a party trick. That’s what the work looks like when the production part stops being the bottleneck.
What I’d tell a founder to budget
Okay, numbers. Mine, for a startup, in 2026. Take them as the shape of the thing, not a menu.
New structure, new words, a homepage that finally says the thing.
- Scope8 to 15 pages
- Time4 to 6 weeks
- Heaviest lineThe deciding + the words
- Lightest lineThe build
THIS IS THE ONE MOST FOUNDERS MEAN
A refresh is the same site with a new skin. Same pages, same words mostly, a new look and a faster build under it. If your structure is right and your site just looks like 2021, this is you. Two to three weeks, and it’s the cheapest thing on this list because almost none of it is deciding.
A redesign is what most founders actually mean. New structure, new words, a homepage that finally says the thing. Eight to fifteen pages, four to six weeks, and the price sits in the middle of the agency range above, not at the bottom, because the deciding and the words are in it. This is the one people mean when they say “redesign.”
A rebuild with product is when the site is the front door to something real: onboarding, booking, payments, an admin desk, a customer login. The site part is the same as a redesign. The product part is its own animal, and it’s the one place where the build is still a serious line, because now it has to be correct, not just pretty. The London practice in my work is one of these, front door to billing.
If a quote comes in way under the middle band for a real redesign, look for what’s missing. Nine times out of ten it’s the words. Someone is going to send you a beautiful template with lorem ipsum in it and call the copy “client-provided,” and you’ll find out what that means in week seven.
How to make it cheaper, honestly
Since the price is mostly decisions, the way to lower it is to make the decisions cheap. Not fewer of them. Cheaper.
Decide who signs off before you start, and make it one person. Every extra approver isn’t extra input, it’s an extra week. Write the words first, or accept that I’ll write them and you’ll edit, but don’t leave it as a box on a plan that says “content.” Ship a smaller site than you want. Fifteen pages nobody finished beat forty pages that launch in Q2. And be in the meeting. To be fair, founders have a company to run, and a site is one of nine things on fire. But an hour a week of you deciding things live is worth more than any discount I could offer, and it’s the only one that also makes the site better.
I’ll probably have to rewrite this one by spring. That’s fine. The numbers will move, the build will get cheaper still, and the part that costs the most will stay exactly where it is. The build’s the easy part now. It’s been the easy part for a while. What a startup website redesign costs in 2026 is mostly what it costs to figure out what you’re saying, and what it takes is mostly you.