A site scoped at eight weeks gets delivered in five months, and the cause is almost never the code. Projects stall on design approvals, on waiting for copy, and on access credentials somebody will "find in a minute". Below is the full route from brief to launch: what happens at each stage, how long it genuinely takes, what the studio owns versus what you own, and the three points where the schedule reliably breaks.

The whole route: seven stages

The sequence below is the same for a landing page and for a store — only the duration of each step changes. The order is not negotiable: starting design without structure, or build without approved layouts, always costs more than it appears to save.

1. 1

   Brief and research — 3–7 days

   A conversation about the business, not the website: who buys, how they decide, who you compete with. Search demand is collected and competitor sites pulled apart in parallel. The output is a document you can argue with before anything is drawn.

2. 2

   Structure and wireframes — 5–10 days

   Page map, queries assigned to pages, greyscale screen schematics. This is where block order and page content get decided. A change here costs minutes; the same change during build costs days.

3. 3

   Design — 10–20 days

   One page first as the stylistic decision, then the remaining templates. Mobile layouts for every screen, not "we will adapt it later". Two revision rounds are built into this stage.

4. 4

   Build and development — 15–30 days

   Layouts become working pages, the CMS goes in, and forms, integrations, filters and payments get written. The longest stage, and the most predictable — provided the earlier ones were done honestly.

5. 5

   Content population — 5–15 days

   Real copy, photography, products, meta tags. Runs in parallel with development if content exists, and sequentially if it does not. This is where a month usually disappears.

6. 6

   Testing — 3–7 days

   Real devices, not just a browser emulator: forms, payments, behaviour on a slow mobile connection, speed checks and structured data validation.

7. 7

   Launch and stabilisation — 2–5 days

   Production domain, SSL, redirects from old addresses, sitemap, analytics and Search Console. Then two weeks of watching errors and real behaviour.

Brief and structure: the stage that decides everything else

The temptation to skip this is enormous — nothing is visible, no layouts exist, and money is already going out. Yet this is where it is settled whether the site can be found at all and whether it holds the pages people need. Rewriting the structure after build means rewriting the build.

• Keywords before structure. Collect real search demand first, then create pages against query groups. The reverse order produces a handsome site nobody searches for.
• Teardown of 3–5 competitors. Not to copy, but to see which sections and arguments have become the baseline expectation in your market.
• Integration list from day one. CRM, payments, shipping, booking calendars. Every integration discovered mid-build adds 5–10 days and reopens the quote.
• A written spec, not verbal agreements. Minimum contents: page list, feature list, who supplies content, how many revision rounds, acceptance criteria. What belongs in it is detailed in our guide to website technical requirements.

Design and approvals: the main bottleneck

The design stage rarely fails because of the designer. It fails because of the approval process. The classic pattern: the layout goes to the director, who forwards it to a partner, who forwards it to their spouse, and ten days later forty comments arrive, half of them contradicting each other.

1. 1Appoint one decision-maker. Gather opinions from anyone you like, but one person says the final yes. This is the cheapest way to remove two weeks from a project.
2. 2Agree the format for feedback. Comments in Figma or a numbered list in a document. Voice notes and screenshots with red arrows leak time on every iteration.
3. 3Review layouts on a phone. That is where most of your visitors will be. Approving a design only on a 27-inch monitor means approving a different website.
4. 4Separate taste from task. "I do not like the blue" is taste — discussable. "The button is invisible against the photo" is a task — fix it. The first can wait, the second cannot.
5. 5Do not add pages during design. Every added page reopens both the timeline and the quote; park them as a second phase after launch.

Build and content: who owns what

This is the longest stretch and the calmest — provided the content exists. The most common deadline failure here is not technical: the site is built and where the copy should be there is Lorem ipsum, because the client was writing it and ran out of time.

Launch and the first two weeks

Launch is not a celebration, it is a technical procedure with a checklist. The biggest risk is migrating from an old site without a redirect map: addresses change, old pages return 404, and organic traffic drops 30–60% for several months. The second most common failure is the site quietly staying out of the index because a noindex from staging travelled into production.

The following two weeks are the stabilisation stage, and it belongs in the contract. These are not "leftovers" — real traffic always surfaces browsers and behaviours testing did not. Record your first speed measurements here too: that is the baseline every later improvement is judged against. On picking a partner who treats this stage as normal rather than as an argument, see our guide to choosing a web agency.

Frequently asked questions

In short

• Seven stages: brief, structure, design, build, content, testing, launch. The order does not change.
• Roughly a third of the calendar time is work on your side — feedback, copy, access, decisions.
• The costliest delays come from design approvals and unfinished content, not from development.
• Appoint one person with the final yes; it shortens projects more than any technology choice.
• Plan two weeks of stabilisation after launch — real traffic always finds what testing missed.