Four to six weeks, brief to live. We put the number on the site because vague timelines are how projects rot, and because we would rather be held to something. Here is what actually happens inside it, week by week, including the parts that are your job rather than ours.
Week zero: casting and the brief
Before the clock starts there is a conversation and then a cast. We read the brief, decide what the project genuinely needs, and assemble the small group that fits it: an identity designer and a writer for a brand-led build, a developer with payment scars for a store doing real volume. This takes a few days, it is not billed as production, and it is the single decision with the largest effect on the six weeks that follow.
The brief itself needs three things from you: what you sell, who you want buying it, and what you have already decided is not up for discussion. That last one saves a fortnight. Founders who tell us on day one that the name is staying and the green is non-negotiable get a much better week two.
Weeks one and two: direction
Two or three directions, presented as real pages rather than mood boards, because nobody can approve a feeling. You see type in position, colour under a product shot, a hero and a product page and a checkout step. We ask for one round of consolidated feedback at the end of week one and a decision at the end of week two.
Consolidated is the operative word. Feedback arriving in six messages over nine days from four people who disagree with each other is the most common way a six-week project becomes a ten-week one.
Weeks two to four: build
Direction is signed, so the build starts while the words are still being written. Templates, components, the store spine: products, variants, tax, Przelewy24, BLIK, Stripe where cards matter, InPost lockers because that is what people here actually choose at checkout. Staging goes up early and stays open. You watch it get built rather than waiting for a reveal.
The dependency in this stretch is yours: product photography and a clean product spreadsheet. We can write every word on the site. We cannot photograph stock that has not been made yet.
Weeks four to five: content and SEO
Copy lands in the real templates, never in a document. Page titles, meta, the structure answer engines read, internal links between collections and journal pieces, product descriptions that sound like a person wrote them because a person edited every one. This runs in parallel with the tail of the build, which is precisely why the calendar works at all.
Week six: launch
Redirect map from the old URLs, analytics, transactional emails tested with real addresses, a live payment through every method you accept, checked on a phone on the network your customers actually use. DNS moves on a weekday morning, never on a Friday afternoon. Then two weeks of aftercare while the first real orders find the edges no test order ever finds.
What makes it slip
Six weeks is not how fast we work. It is how fast we work together.
Four things cause almost every overrun we have seen, and three of them are preventable before anyone opens a design tool:
- No single decision maker. Name one person who can say yes. Committees do not approve, they average.
- Content arriving last. Photography, product data and legal pages should be in motion during week one, not commissioned in week five.
- Scope added mid-build. A booking system, a second language, a loyalty scheme. All reasonable, all a separate phase, none of them free of the calendar.
- Third parties. Payment verification, a distributor's stock feed, a domain held by a former developer who is not replying. We chase these from day one because they are the only category we cannot compress.
When a project does run to eight weeks, it is almost never the design that took longer. It is a decision that sat waiting in somebody's inbox. That is worth knowing before you brief anyone, ours or otherwise, because the half of the timeline you control is the half that usually moves.