How a website gets built in 12 working hours

Ask how long a website takes to build and the answer comes back in weeks. Four to eight is the common quote for a custom small-business site. Plenty of firms quote three months.

My builds take about 12 working hours.

Both numbers are accurate, and they measure different things. The first one measures the calendar, from kickoff to launch, counting every day the project sits waiting for something. The second one measures hands on the work.

The hours are short because the decisions were made before the build started.

By the time the first page session opens, the words exist and the colors exist. The page structure exists. One person approves the result.

What is left is assembly.

Where the 12 working hours go

The build runs in a fixed order, and the order is what keeps it short.

The build starts with a design system, made once for the brand and used by every page after it. The inputs come from the brand's visual identity guide: exact hex values, licensed typeface files, the logo set.

That step takes about 10 minutes. It is the reason no page later in the build reopens a question about color or type.

The skeleton comes next: navigation, footer, the newsletter block, and the closing section that invites a consultation. Each one gets built as a reusable component and reviewed at phone width before desktop. Nothing else gets built until those four are approved.

Then come the pages, one page per working session, nine in all. Five of them carry the argument: home, method, work, about, and consultation. The other four are the journal index, the journal post template, privacy, and the 404.

Each session starts with a finished copy document, and the copy goes in as written.

The largest share of the 12 hours sits here, in composition. One working designer publishes the averages he has gathered from his own projects and from other designers. A landing page runs around 10 hours of design time, and a five-page site around 25.

Set against that, 12 hours for nine pages plus the deployment is fast. It is possible because composition starts from settled decisions.

A correction pass follows the pages, run in code rather than in the design tool. Every line of placed copy gets pulled out of the built pages and checked against the brand guides. Anything that drifted comes back as a numbered list of edits, applied in one pass.

The last stretch is deployment. That covers the repository, the journal posts, and the redirects from the old site. It also covers the consultation form, the schema markup, and the live hosting.

The hours collapse because the questions were asked first

Three hours of recorded interviews happen before the build starts. They cover the business, the services, and the reasoning behind how the client works. They also draw out everything the client knows about their customers.

Those transcripts become a suite of brand guides that settle voice, audience, services, and visual standards in writing.

By the time the build starts, the two things that stall other website projects are finished.

The first one is content. Ask agencies what breaks their timelines. The answer is the copy, waiting on a client who ran out of hours.

The pattern repeats. Wireframes get approved, then the design. Then weeks of nothing, while someone finds time to write the pages on top of their day job.

One copywriter's task-by-task breakdown puts at least 11 days of professional time into a thorough 20-page site. That is a fair picture of what the calendar absorbs while nobody is building.

Here the writing comes out of the captured material. It is produced through the trained system and reviewed before composition starts.

The second one is approval. Website work moves in stages, and each stage waits for a review. Designs and copy sit untouched for days while feedback trickles in from several stakeholders.

One client decides here, and that client has already read their own thinking in the copy. Reading your own reasoning back is faster than judging a stranger's guess at your business.

The calendar is a separate number, and it belongs to you

Working time is short. The calendar is set by how quickly the interviews get done.

Complete the capture inside a week and the launch lands in about two. Give one hour a week and the same engagement stretches across six weeks or more.

Building does not start until the capture is complete. Nothing is lost at the slower pace except time.

Your launch date depends on your calendar, not mine.

What the 12 working hours do not include

The 12 hours cover the site itself, from the design system to the live page. They do not cover the three hours of capture, or the guide suite written from those transcripts. They do not cover the hour that closes the gaps building surfaces.

Photography sits outside them as well. The build supplies a photographer brief and a shot list, and the shoot is a separate engagement.

Speed is only safe when someone checks the parts these tools skip. One consultant audited five business websites built with AI tools in April 2026. Most of them shared the same gaps.

The headings read well and signaled nothing to a search engine. Robots files were missing or misconfigured, sitemaps were broken, and schema was absent.

Five sites is a small sample, and she says so in the piece. Her findings work as a checklist rather than as a measurement.

Designers building client sites this way report the same class of problem from the inside. They name layout drift, responsive states missing on mobile, and pages that look finished and are wrong underneath.

The correction pass and the deployment stretch exist for that reason. A page can read well and still fail on the parts nobody sees. That page is not finished, and the only way to know is to go and look.

One more limit belongs here. The first engagement built this way ran past 12 hours. The method was being written at the same time.

The figure is what the same work costs now that the order is settled.

The 12 hours are what is left after the deciding

A website takes 12 working hours when nothing is being decided during the build. The deciding happened earlier, in recorded interviews and in a set of guides. It took longer than the build did.

Move the thinking to the front and the hours at the back get short.