In shortA quote is a price attached to a set of assumptions. When two quotes differ, the assumptions differ — so compare those, not the totals.
The spread is real, and it is mostly not markup.
It is genuinely normal to receive quotes for the same redesign that differ by a factor of five. Most business owners read that spread as evidence that somebody is overcharging, and occasionally somebody is. Far more often the quotes are pricing different jobs.
"Redesign our website" is not a specification. It can mean applying a new visual theme to the existing structure, or it can mean rebuilding the information architecture, rewriting every page, migrating a decade of content and integrating a booking system. Those are different amounts of work by an order of magnitude, and both are honest readings of the same sentence.
The variables below are what move the number. None of them are visible in a finished website, which is exactly why they are so easy to leave undecided until they turn into a change request.
What actually drives the price.
Roughly in order of how much they move the total.
How many pages there really are
Almost everyone undercounts. A business that describes itself as having an eight-page site often has forty once you include every service page behind a dropdown, every location page, old landing pages still receiving traffic, and the blog posts nobody has looked at since 2018. Ask any agency what page count they quoted against — if the answer is your estimate rather than a crawl of the site, the quote is built on a number nobody checked.
Whether the content exists
This is the biggest hidden variable and the most common reason projects run late. Reorganizing copy that already exists is quick. Writing service pages that never existed is a different discipline being billed at a different rate, and waiting on text the client intends to write themselves is what most stalled projects are actually waiting on. Decide who is writing, and price that.
Whether the old content has to move
Migrating hundreds of pages from an old CMS is mechanical, tedious, and rarely fully automatable — old content is full of inline styles, broken embeds and links to pages that no longer exist. It is usually priced per page, and it is worth asking whether every page genuinely needs to come across.
What has to connect to what
A brochure site with a contact form sits at the bottom of any range. Booking systems, payment providers, quoting tools, CRMs and inventory feeds each add integration work, testing, and a dependency that can break independently of the site. Every integration is a small project attached to the large one.
How much search equity is at stake
A new site with no traffic can change every URL freely. A site with hundreds of indexed pages and years of accumulated links needs an inventory, a mapped redirect plan and post-launch monitoring. That is real work, it is skilled work, and a quote that does not mention it has either absorbed it silently or is not doing it.
Custom design versus a template
A template constrains the design to what the template does, which is often perfectly adequate and considerably cheaper. Custom design costs more and is worth it when the business has a real brand to express or a structure no template anticipates. Neither is the wrong answer; being unclear about which you are buying is.
Who edits it afterward
A site the owner can update needs a CMS, an editing model that survives contact with a non-technical user, and a handover. A site only the agency can change is cheaper to build and more expensive to live with.
What a suspiciously cheap quote usually leaves out.
Cheap is not automatically bad — a small site on a template genuinely is inexpensive. But a very low number for a large site usually means one of these is missing.
- Content migration, quoted separately or assumed to be your job
- Copywriting, with "client to supply all text" buried in the terms
- A URL inventory and redirect plan, so existing search traffic is on its own
- Mobile layouts designed rather than inherited from the desktop design
- Accessibility, which is a legal exposure in several markets and a real cost to retrofit
- Performance work, so the site is fast on the developer machine and slow on a phone
- Testing across real devices rather than a browser window resized
- Any post-launch fixes, so the first bug becomes a new invoice
- Ownership of the result — check whether you can take the site elsewhere
What an expensive quote should be buying.
Higher is fine. Higher for no articulable reason is not.
- Discovery that produced findings you had not already told them
- Information architecture decided from evidence rather than from taste
- Original writing, not your existing copy pasted into a new layout
- Design work that is specific to your business rather than adapted from a demo
- A migration plan that names your actual URLs
- Performance and accessibility as part of the build rather than a later pass
- A named person who will still be reachable in six months
How to compare quotes fairly.
Normalize the page count first
Get every agency quoting against the same number, from an actual crawl. Until that is true, the totals are not comparable and no amount of staring at them will help.
Ask each one what they assumed about content
This single question explains most of the spread. An agency assuming you supply all copy and one assuming they write it are quoting jobs that differ by weeks.
Ask what happens to your existing URLs
A team that answers this fluently has done migrations before. A team that has not thought about it is a risk you are taking on unpriced.
Ask what is explicitly out of scope
The exclusions are more informative than the inclusions, and a good agency will have them written down. Vagueness here reliably becomes a change request later.
Compare payment structure, not just totals
Staged payments tied to real milestones protect both sides. A large deposit with no defined stages is a risk regardless of the number.
The question worth asking instead.
"What does a redesign cost" is the wrong first question, because the answer is genuinely "it depends" and nobody can shortcut that honestly. The more useful question is what the current site is costing you — in enquiries that do not arrive, in calls that go to a competitor whose number was tappable, in the search traffic that a site built for desktop no longer earns.
That number is also hard to pin down, but it is at least a number about your business rather than about the market. A business losing a few jobs a month to a broken contact form has a very different calculation from one whose site is merely unfashionable.
Which is why any serious agency should look at your actual site before quoting, and why a fixed price offered before anyone has seen it is a price attached to assumptions rather than to your project.