The honest answer is that it depends, and the unhelpful thing about that answer is that it is usually where the conversation stops. So here is what it depends on, with real timelines from recent builds, so you can place your own project on the scale before you ask anyone for a quote.
The short answer, by type of site
- A focused business site (five to fifteen pages, a clear brand, content mostly ready): four to eight weeks.
- A full bespoke site with SEO architecture (service pages, area pages, case studies, an article library, calculators or configurators): six to fourteen weeks.
- A platform (customer accounts, portals, commerce, integrations with other systems): twelve weeks and up, usually phased.
- Anything with a large content migration running alongside: add the time it takes to audit and move the content, which is often the longest single task.
Those ranges are typical rather than promised. What follows is what moves a project within them.
Real timelines from recent work
- Milkov & Son Construction: six weeks from audit to launch. A full rebuild on Next.js with thirteen service pages, twelve area pages, a nine-step project cost calculator and a migrated library of over two hundred articles, every legacy URL preserved or redirected.
- CC Finance: six weeks from kick-off to launch for a new financial-services website.
- OuterWings: eight weeks for an e-commerce build.
- Petras Architecture: twelve weeks from kick-off to launch.
- Glass Openings: eight months for a site built around a 3D product configurator, which is a product as much as a website.
The spread is the point. The six-week projects were not smaller businesses; they were projects where the content existed, the decisions were made on time and the scope was fixed at the start.
What the weeks are spent on
Discovery and architecture: one to two weeks
What the business sells, how its customers decide, which searches the site has to win, which existing URLs have to survive. The sitemap and page types come out of this, and so does most of the SEO. Skipping it is the most expensive way to save a week.
Design: two to four weeks
The visual system, then every page type designed in full and approved before any code is written. Rounds of feedback are where this stretches, so agree the number of rounds up front.
Build: two to six weeks
The design becomes code. On a modern framework this runs in parallel with content and integrations. The longest builds are the ones with the most bespoke functionality: calculators, portals, anything that talks to another system.
Content and migration: alongside everything, and often the critical path
Writing new pages, moving old ones, fixing what was wrong with them, mapping every legacy URL. On sites with a history this is frequently the task that decides the launch date, not the build.
QA, SEO checks and launch: one week
Every page tested on real devices, every redirect verified, structured data validated, analytics events firing, DNS moved. A launch day is a checklist, not an event.
What makes a project slower
- Content that is not ready. The single most common cause. A site cannot be finished before the words are.
- Approval bottlenecks. One decision-maker who reviews fortnightly adds a fortnight to every round.
- Scope that grows in the middle. Every good idea in week five costs more than the same idea in week one.
- Integrations with systems nobody documented. The CRM, the booking tool, the payment provider. Find out early what they can actually do.
- Photography and brand work happening at the same time. Sequence them before the design phase, or the design gets done twice.
What makes it faster
Content ready before design starts. One accountable decision-maker with a fixed review rhythm. A scope that is written down. A framework and a content system that let design, build and content run in parallel rather than in sequence. And a discovery phase that was taken seriously, because most mid-project slippage is a discovery question that was never asked.
Fast is not the same as rushed
Six weeks is fast for a full bespoke build. It was possible on the projects above because the discovery was done properly, the content was audited before it was moved, and the client made decisions on time. It was not possible because corners were cut. A site launched a fortnight early with an unfinished redirect map will spend the next six months recovering rankings it did not need to lose.
If an agency promises a full bespoke build in a fortnight, ask which of the phases above they are skipping.
What a six-week plan actually looks like
For a full business site where the content exists and one person can sign things off, the plan that has worked for us runs like this.
- Week 1: discovery. Audit the existing site URL by URL, verify every claim the new site will make, agree the sitemap, the page types and the searches each page has to win. Content gaps are listed and assigned.
- Week 2: design system and key pages. Typography, colour, components, then the homepage and the primary service page designed in full. First approval.
- Week 3: remaining page types and build start. Area pages, case studies, article templates and any tool designed while the build of approved pages begins. Content migration starts in parallel.
- Week 4: build. Every page type built, the content system wired up, integrations connected, the tool or calculator functional. Client reviews real pages, not mock-ups.
- Week 5: content in, redirects mapped. Every page populated, every legacy URL mapped and tested, analytics events defined and firing on staging.
- Week 6: QA and launch. Device testing, accessibility and performance checks, structured data validated, DNS moved, sitemap submitted, redirects verified on the live domain.
Two things make that schedule hold: content that arrives in week three rather than week five, and decisions that come back within two working days. Remove either and the same project takes ten weeks, not because anyone worked slower but because the work waited.
How to brief for a realistic timeline
If you want an accurate estimate rather than a hopeful one, give the agency four things before asking for a date: the list of pages you expect, the state of the content for each (written, needs editing, does not exist), every system the site has to talk to, and the name of the person who will approve the work. An agency that quotes a date without those four has quoted a guess, and it is usually an optimistic one.
Most slippage is a discovery question that was never asked.
Michail leads strategy, development, and growth at Molo Agency. He works closely with clients on SEO, web design, automation, and custom software projects, with a strong focus on building systems that generate real business growth.