PMI's 2018 Pulse of the Profession found that 52% of projects completed in the previous year experienced scope creep or uncontrolled changes to scope, up from 43% five years earlier. A four-week website build does not survive that rate by working faster. It survives by fixing the date and letting the scope move, at a small number of named points where someone with authority decides.
This guide is the full version of the four-week plan we sketched in The B2B SaaS Website Playbook. It covers what fits in four weeks, the inputs that must exist before day one, the four decision gates, and the launch checks, with the published guidance from Google, Webflow, W3C and others behind each one.
What fits in four weeks, and what does not
Four weeks is an appetite, not an estimate. Shape Up describes appetite as "a time budget for a standard team size" and pairs it with the principle it calls "fixed time, variable scope". Its small-batch projects take one designer and one or two programmers one or two weeks; its big batches take the same team six weeks. A four-week marketing site sits between the two, with a similar team: one strategist or designer, one or two Webflow developers, and a client-side owner.
Our own site claims we ship a typical marketing site in three to four weeks, set against the two months it says a proposal-heavy agency quotes. That is our marketing, so treat it as a vendor claim. The useful part is the word typical, and what it excludes.
| Item | Fits in four weeks | Moves the date or needs a second phase |
|---|---|---|
| Static pages | Up to about a dozen page types | Dozens of bespoke one-off layouts |
| Design | Three or four key layouts plus a component set | Every page designed separately |
| CMS | Three to five Collections, such as posts, case studies, team | Large migrations, multi-reference data models |
| Content | Supplied or approved by the client, with one owner | Copy written from scratch for every page |
| Integrations | Forms to a CRM, analytics, scheduling embed | Gated portals, custom APIs, member areas |
| Locales | One language | Several locales with localised slugs |
| Redirects | A mapped list, wildcards where possible | Thousands of legacy URLs with no inventory |
Source: our recommendation, based on Octo Agency's four-week marketing-site process; Webflow redirect guidance from the Webflow Help Center
The right column is not a list of things we refuse. It is a list of things that need their own appetite. Our white-label builds for Web Tonic run two to three weeks per site across more than 20 Webflow sites in twelve months, and the largest carried a CMS of over 1,200 items. That CMS took planning in the first week, not more weeks at the end.
Inputs that must exist before day one
Most four-week builds are lost before they start, in the gap between signing and the first working session. We would ask for six inputs before the clock begins, and start the clock only when they arrive.
| Input | Owner on the client side | Why it is needed before day one |
|---|---|---|
| One named decision owner | Sponsor | Gates need one person who can say yes or no |
| Access: domain registrar or DNS, Search Console, GA4, CRM | Marketing ops or IT | DNS and tracking changes in week four depend on it |
| Existing URL list or crawl | Marketing | Redirect mapping starts in week one |
| Brand assets: logo files, fonts, colour values | Brand or marketing | Design in week two cannot start from screenshots |
| Content inventory: what exists, what is new, who writes it | Marketing | Sets the content freeze date in week three |
| Approval list: legal, compliance, customer sign-offs | Sponsor | Case studies and security pages wait on these |
Source: our recommendation; the DNS, redirect and staging requirements come from Webflow and Google guidance cited below
The decision owner matters most. A committee can advise, but a gate only works when one person signs it. Our companion post, B2B Website Strategy: How to Write the Page Map Before Anyone Opens Figma, covers the strategy document this person signs at the first gate.
The four decision gates
A gate is a point where the project stops, the decision owner looks at a defined piece of evidence, and the answer is yes, or no with a pre-agreed consequence. Shape Up describes the same idea as a circuit breaker: "if the work doesn't get done, the project doesn't happen", and extensions are the rare exception rather than the default. On a client site, cancelling is rarely the right answer. Moving scope to a second phase usually is.
| Gate | When | The question | Evidence on the table | If the answer is no |
|---|---|---|---|---|
| 1. Scope and message | End of week 1 | Do we agree what we are building and what it says? | Page map, template list, Collections, message document, redirect map started | Cut pages to a phase-two list, or move the date now |
| 2. Design | End of week 2 | Do these layouts and components cover every page? | Three or four key layouts, component set, mobile views | One revision round; pages not covered move to phase two |
| 3. Content freeze | End of week 3 | Is every launch page on staging with real, approved content? | Staging site with real content, open-items list | Missing pages launch later; the date holds |
| 4. Go or no-go | Mid week 4 | Does the site pass the launch checklist? | QA record, redirect test, tracking test, accessibility scan | Fix blockers only; non-blockers go to a post-launch list |
Source: our recommendation; the circuit breaker and fixed-time principle come from Basecamp's Shape Up
Week one: strategy, ending at gate one
Our process starts with a Strategy Week, and our site says so in plain terms: "Every project starts with a Strategy Week." What happens inside it is detailed in the strategy post and in the B2B SaaS Website Playbook, so we will not repeat the page map here.
For the four-week plan, three outputs of week one matter more than the rest. The template list. Every page on the map is assigned to a template, and the template count, not the page count, sets the build effort. The Collection model. Our Webflow CMS Collections for a B2B Site post explains the model; the point here is that changing a Collection's fields after content goes in costs days. The redirect map. Google's site-move guidance says "it's important to map your old site's URLs to the URLs for the new site". Starting the map in week one means launch week is spent testing redirects, not writing them.
The analytics plan is the fourth output, and our post The Analytics Stack for a Four-Week Site Build covers its week-by-week track.
Week two: design, ending at gate two
We would design three or four representative pages, typically the homepage, a service or product page, a case study and a conversion page, plus the component set they share. Every other page is assembled from those components in week three.
A short usability check fits here. Jakob Nielsen's article for Nielsen Norman Group, from 2000 but still the reference, found that a first study with five participants found 85% of the usability problems, and recommended spreading budget across many small tests. On a four-week plan, that means five sessions on the clickable homepage and conversion path on day eight or nine, not a study after launch.
Gate two asks one question: do these layouts and components cover every page on the map? If a page needs a layout nobody designed, it either reuses the nearest one or moves to phase two. We would allow one revision round on the key layouts. A second round is a scope decision for the decision owner, made at the gate, with the date in view.
Week three: build and content, ending at gate three
Templates, components and Collections go up on the Webflow staging subdomain. Webflow describes the staging subdomain as useful "to test custom code and get approval on your site's design before pushing your site live", and its indexing can be switched off so the draft does not appear in search.
Real content goes in this week, not placeholder text. This is where four-week projects slip, and the cause is rarely the build. It is the case study waiting for a customer's legal team, the pricing page waiting for a finance decision, or the security page waiting for a compliance lead. All three should have been on the approval list before day one.
Gate three is the content freeze. Every page that will launch is on staging with approved content. Anything not there goes to the phase-two list with an owner and a date. This is the gate that protects launch, and it is the one most often negotiated away. We would hold it.
Week four: QA, go or no-go, launch
The first half of week four is QA on staging. The second half is launch and handover. The go or no-go decision sits between them, and it is made against a list agreed in week one.
| Check | Pass condition | Published basis |
|---|---|---|
| Redirects | Every mapped old URL returns a 301 to its new URL | Google recommends permanent redirects (301 or 308), kept generally at least a year |
| Redirect volume | Under about 1,000 rules, wildcards where possible | Webflow recommends 1,000 as a best-practice maximum |
| Localised paths | Each locale path has its own rule | Webflow redirects do not apply to localised slugs automatically |
| Domain | Site plan active; DNS records set early in the week | Webflow: DNS can take up to 48 hours, usually a few hours |
| Speed, lab check | LCP within 2.5 s, INP 200 ms or less, CLS 0.1 or less on key templates | Google's Core Web Vitals thresholds, measured at the 75th percentile |
| Accessibility | No contrast, alt text, form label or empty link failures on key templates | WCAG 2.2; these four lead the WebAIM Million failure list |
| Forms and tracking | Every form submits, notifies and fires its event | Our recommendation |
Source: Google Search Central, Webflow Help Center, web.dev, W3C, WebAIM
Speed. The Core Web Vitals thresholds are measured on real-user data at the 75th percentile. The Chrome UX Report is a 28-day rolling average, and pages below a popularity threshold are not included at all, so a new site has no field data on launch day. Run lab checks before launch and read the field numbers a month later.
Accessibility. WebAIM's February 2026 scan of a million home pages found low-contrast text on 83.9%, images without alt text on 53.1%, form inputs without labels on 51% and empty links on 46.3%. All four are cheap to fix in week four and expensive to find after launch.
Timing. Webflow requires a Site plan before a custom domain connects, and it notes that DNS changes may take up to 48 hours to propagate, though usually within a few hours. We would set DNS early in week four, launch mid-week, and keep the final day for handover. After a move, Google warns to "expect temporary fluctuation in site ranking" and says a medium-sized site can take a few weeks or more to show its new URLs, so launch-week traffic is not the verdict.
Handover follows launch. The Webflow Client Handover Kit post lists what should change hands and the twelve settings to check.
Where four-week builds slip, and which gate catches it
The slips are predictable, which is why the gates are where they are.
The absent decision owner. Sign-offs by committee turn a two-day review into a week. Gate one fails if the owner has not signed, and the date moves then, while it is cheap.
The late page. A page added in week three has no design, no content and no template. It goes to phase two by default.
The approval queue. Customer quotes, logos and compliance text depend on people outside the project, so we would plan for them to arrive late. Gate three catches them, and the pre-agreed answer is to launch without them.
The migration surprise. An old blog with hundreds of posts and no URL list turns redirect mapping into a project of its own. The pre-flight crawl finds it before day one.
None of these is a build problem. A four-week Webflow build is mostly a decision schedule with a build attached, which is also why the agency timelines collected in our Website Redesign Statistics 2026 post run longer, six to ten weeks for one B2B example.
What this means for planning your build
Fix the date and write down what can move. Name one decision owner, collect the six inputs before day one, and agree the four gates and their consequences in the first meeting. Design four pages, freeze content at the end of week three, and launch against a checklist with published thresholds rather than a feeling. Anything that misses a gate goes to a dated phase-two list instead of pushing the launch.
Our Webflow development service page explains how we scope a four-week build.
