Everything here is free and needs no account. We built each one because a job needed it and nothing good enough existed.

All tools
Article

Fixed price or bands: how to read a web design quote

Rolled and flat engineering drawings covered in dimensions and tolerance callouts, tinted blue

Three quotes are on your desk. One is a single number. One is a range with a page of assumptions attached. One is low enough that you already suspect something has been left out.

The single number looks like the safest of the three. It is the one you can forward to a director without a covering note. It is also, more often than not, the one that moves.

What a web design quote is actually pricing

A web design quote is not priced by page count, and it is not really priced by hours either. It is priced against a scope: a set of assumptions about what the site has to do, who has to be convinced by it, what already exists in usable form, and how many decisions are still open.

Change any one of those and the number changes with it. A five-page site for one audience with finished copy and a photo library is a different job from a five-page site for three audiences where the copy has to be written, the certifications have to be gathered from a filing cabinet, and two directors disagree about who the site is for. Same page count. Not the same project.

So the useful question when you are reading a quote is not “how much”. It is “what does this document assume, and how confident is the person who wrote it”.

Why a fixed price against a loose brief is a bet

Engineering and construction already solved this problem, and solved it formally.

AACE International’s cost estimate classification system ties the accuracy an estimate can credibly claim to how well the project is defined at the time of estimating. At the earliest stage, with roughly nothing defined, an estimate carries an accuracy range of about minus 50% to plus 100%. Only at the top class, with the project substantially engineered, does the range tighten to roughly minus 10% to plus 15%. The classification is explicit that the range is part of the deliverable, not a caveat bolted on afterwards. Presenting a screening estimate as a funding figure is treated as a misuse of the method.

Nobody in a QS office finds that controversial. Web projects are smaller and less dangerous, but the logic does not change with the size of the number.

When a studio quotes one fixed figure against a brief that is a paragraph long, the uncertainty has not disappeared. It has been priced by someone, and only two things can have happened. Either they padded the figure to cover the range, in which case you are paying for scope you may never use. Or they did not, in which case the shortfall gets recovered later: through change requests on everything that was not written down, through a junior doing the work instead of the person you met, or through the parts of the project that are invisible in a demo, like structure, writing and testing.

That is the honest description of a fixed price for undefined scope. It is a bet, and one side has to win it.

A band does something different. It says: here is what we think this is, here is what would push it to the top of the range, here is what would keep it at the bottom, and we will fix a price once the scope is agreed. The uncertainty is on the page instead of inside someone’s margin.

What the exclusions tell you

Most people read a quote from the top. The informative part is usually near the bottom, in the section written in smaller type.

Look for whether the following are named at all, either as included or excluded:

  • Copywriting. Who writes the words. “Client to supply content” is a legitimate position and a common one, but it is also the single most frequent cause of a project stalling for two months.
  • Photography. Whether a shoot is in scope, or whether the site will be built on stock and whatever is on the shared drive.
  • Templates versus pages. Ten pages built from three templates is a smaller job than five pages each designed individually. A quote that counts pages and not templates has not been thought through.
  • Revisions. Not just how many, but what reopens a phase and what that does to the launch date.
  • Third-party costs. Licences, plugins, fonts, stock, form or booking tools. These are small individually and they are yours to renew forever.
  • Migration. Whether existing pages, posts and files are moved across, and by whom.
  • Support after launch. Whether there is a window, how long it runs, and what counts as a bug rather than a new feature.
  • Ownership. What transfers to you at handover: files, logins, hosting account, domain, source.

A quote with no exclusions is not a simpler quote. It is a less finished one. Every project has a boundary, and a document that does not draw it has either not found it yet or has decided not to show you where it is.

The same applies to what a studio does with pricing publicly. Nielsen Norman Group has been reporting for years that business buyers rank price as their top information need and routinely leave sites that hide it. Quote on request is a choice about who gets to know the number first, and it is rarely made in the buyer’s interest.

How do you compare two quotes that look nothing alike?

Do not start with the totals. Start by making the two documents describe the same project.

Write out a single scope list of your own: number of templates, number of audiences to separate, who writes copy, who supplies images, integrations, languages, migration, support window, ownership. Send it to both and ask each to price against it. Half the difference between two quotes usually evaporates at this step, because you find one of them was never including copywriting and the other always was.

Then compare three things.

The assumptions. Ask each studio what would move the number, in which direction, and by roughly how much. A studio that can answer that quickly has priced this kind of work before. A studio that says the price is the price has either done very little thinking or is not planning to share it.

The process boundary. When is scope agreed, when is structure signed off, and what happens to the timeline when feedback arrives after a phase has closed. This is where projects actually go over, not in the design.

The recurring cost. Hosting, maintenance, licences, domain. A cheap build with an expensive tail is not a cheap build. Our hosting starts from S$200 a year and maintenance from S$200 a month, and both are quoted against what the site is actually doing rather than as a fixed line item.

What we publish, and why

Creatif Work publishes pricing bands rather than quoting on request. Projects start from S$4,000. There are three bands, set by what the project has to do rather than by how many pages it has, and most work lands in the middle one. If a budget sits below the first band we say so on the first call.

The band is not the price. Scope is agreed first, in writing, and the price is set against that scope before design starts. Nothing appears on an invoice that was not agreed beforehand. What moves the number is consistent and easy to state: how many audiences the site has to separate, whether copy and photography already exist, whether certification and prequalification content has to be gathered from scratch, and whether anything has to integrate with a system you already run.

Terms follow the same principle. Smaller projects run half on signing and half at launch. Larger ones split 25% on signing, 35% at design sign-off and 40% at launch, with 30 days to pay on the last two so a finance team is not the reason a launch slips. Every build includes 30 days of support after handover for bug fixes, content edits and questions, extended to 45 days on larger builds.

None of that is generous. It is just written down, which turns out to be the differentiator.

The one question worth asking

If you only ask a studio one thing before signing, ask this: what is the most likely reason this project ends up costing more than your quote?

An honest answer arrives immediately, because the person has watched it happen. It will be content, or a stakeholder who was not in the room, or an integration nobody scoped. A studio that cannot name a failure mode has either not run enough projects or is not going to tell you about the ones it has.

That answer tells you more than the total does. When we rebuilt THK Engineering, the scope question was never how many pages, it was whether a main contractor could confirm in one pass that five M&E disciplines were carried in house rather than sublet. Getting that right is what the money buys. Page count was an output, not an input.

If you want the wider picture on what these projects cost in this market, we set that out separately in what a B2B website actually costs in Singapore. Every figure there, and here, is a starting point.