Ask five agencies to quote the same website and you will get five numbers that look unrelated to each other. That is not because four of them are lying. It is because “website” describes anything from a one-page template swap to a bespoke application with user accounts, payments and a permissions model.
Until you know which of those you are buying, comparing prices is meaningless. So before you compare quotes, it is worth understanding what actually moves the number.
Why quotes vary so wildly
A quote is an estimate of hours multiplied by a rate, plus risk. The rate varies less than people assume. What varies enormously is the hours — and the hours depend almost entirely on how much of the work already exists before the project starts.
If you arrive with a finished brand, written copy, and a clear feature list, a competent team can move fast. If you arrive with a rough idea and expect the agency to decide what the site says, what it looks like, and what it does, that is three disciplines of work before anyone writes code.
The cheapest quote is usually the one where somebody has assumed you will supply things you have not thought about yet.
The three things that drive price
How much of the design is original. Adapting an existing theme is hours. Designing a system from scratch — type scale, colour, components, states, responsive behaviour — is weeks. Both produce a website. They are not the same product.
How much of the functionality is custom. A contact form is trivial. A booking system that checks availability across multiple practitioners, takes a deposit and emails an intake form is a small application. The word “form” appears in both descriptions.
How much content someone has to produce. This is the one that surprises people most often. Copywriting, photography and data migration are real costs, and if you do not budget for them they become the reason the project stalls for two months at 90% complete.
If a quote does not state who is writing the content, assume it is you. That assumption is behind a large share of projects that run late.
Where the money actually goes
On a typical business website, the split lands roughly like this: design takes about a third, build takes about half, and the remainder covers content, testing and launch. Project management is folded into those rather than sitting on top.
When a quote comes in dramatically below the others, something on that list is not being done. In our experience it is almost always testing and content — the two items a client cannot easily verify from a proposal document.
What you can safely cut
- Custom animation. Enjoyable to build, rarely changes a conversion rate.
- Bespoke photography for a first launch. Good stock plus a strong layout carries you until there is something real to photograph.
- A blog you have no intention of writing. An empty blog is worse than no blog.
- Page count. Six good pages outperform twenty thin ones, for users and for search.
What you should not cut is performance work, accessibility and analytics setup. All three are far more expensive to retrofit than to build in, and the last one is how you find out whether any of the rest worked.
Questions worth asking any agency
- Who owns the code when we are finished? If the answer is not “you do”, understand exactly what you are renting.
- What happens if we need a change mid-project? Look for a written change process, not reassurance.
- What is included after launch? Thirty days of fixes is a reasonable baseline.
- Who is actually doing the work? The person pitching is often not the person building.
Vague answers to those four predict a difficult project far more reliably than the price does. A higher quote with clear answers is usually cheaper by the end.