When we send out a proposal for a custom website, the timeline section is usually the part clients re-read twice. The quotes they have already received range from "six weeks" to "six months", and there is rarely a clear reason for the spread. Two agencies, the same scope, four months apart on the calendar.
Most of the spread is honest. The agencies are quoting different processes, different team structures, and different definitions of "done". Here is the breakdown we use at Taqwa Tech for a typical Tier 3 marketing site — and the parts of the calendar that nobody mentions until you are inside the project.
The honest summary
A custom marketing site at our studio takes 6 to 10 weeks from kickoff to launch, depending on scope. A custom web application takes 10 to 16 weeks. Both ranges assume the brief is locked at kickoff and that one person on the client side owns the engagement. If either is missing, add 30% to the calendar — and that is being generous.
The work breaks into four phases. None of them are optional. The variable is how much calendar each one consumes.
The four phases
| Discovery | Design | Build | Launch | |
|---|---|---|---|---|
| Typical length | 1–2 weeks | 2–3 weeks | 3–5 weeks | 1 week |
| Client effort | Heavy (5–8h) | Moderate (3–5h/week) | Light (2–3h/week) | Heavy (4–6h) |
| What ships | Brief, sitemap, content inventory | Hi-fi designs, design system, copy direction | Coded, content-loaded, QA-passed site | Live site, redirects, analytics |
| What slips it | Unclear goals, missing stakeholders | Indecision, design-by-committee | Mid-build scope changes | Late content, missing assets |
Each phase has a clear definition of "done" that the next phase depends on. Skipping ahead to compress the calendar — building before design is locked, designing before discovery is done — is the single most common reason projects ship late.
Discovery (1–2 weeks)
Discovery is where most agencies under-invest, and where most projects later slip. The deliverables are unglamorous: a sitemap, a content inventory, a one-page brief, a list of integrations, a list of decisions still open. The point is that every later phase is cheap when discovery is real.
A two-day discovery sprint produces a project that argues for six months. A two-week discovery produces a project that ships in eight weeks.
Design (2–3 weeks)
Design is where the most calendar evaporates if the client side is not aligned. A focused two-week design phase has roughly three feedback rounds — initial direction, refined direction, locked direction. Each round is two business days of agency work and two of client review. If review takes a week, design takes six weeks. The math is unforgiving.
We deliver design in figma, locked section by section. The build starts on the locked sections while later sections are still in design — that is the overlap that compresses calendar without compressing work.
Build (3–5 weeks)
Build is the most predictable phase. A senior engineer working on a locked design knows their velocity within a day. The variability in this phase comes from integrations: a CRM hook with a stable API takes a day; a custom workflow into someone's internal tool takes a week and a half.
Launch (1 week)
Launch looks like a single day on the calendar but consumes a real week. There is content loading, redirect mapping, DNS propagation, analytics wiring, schema validation, accessibility passes, and the final QA round on production. Most of this is invisible until something goes wrong — which is exactly when you want a week of buffer, not an afternoon.
Why "rush" projects cost more, not less
Clients sometimes ask for the calendar to be cut in half by adding engineers. It almost never works that way for marketing sites. Two reasons:
- Design has a serial floor. A site needs to feel like one thing. Three designers working in parallel produce three sites stitched together, not one site faster.
- Coordination tax compounds. A two-engineer build has roughly 1.7× the velocity of a one-engineer build. A four-engineer build has roughly 2.4×. The compression is real but the curve flattens fast.
What rush projects actually buy you is the option to ship earlier — and the cost is paid in two places: in the agency's calendar (we de-prioritise other work, which gets billed) and in your team's calendar (faster feedback rounds mean tighter availability windows). Most of the time, the honest conversation is whether the launch date is real, not whether the timeline can be compressed.
The fastest way to ship a custom site in eight weeks is to plan for ten. The fastest way to ship in twelve is to ask for six.
What actually slips a timeline
After enough projects, the pattern is consistent. In order of damage:
- Late content. Copy, photography, video, case studies. We can build a site without final content, but the final 10% of the build cannot land until the content does. Two weeks of late content is two weeks of late launch.
- Unbooked stakeholders. A founder who is travelling during design week. A CMO whose review comes at week six instead of week two. Every late stakeholder costs roughly the calendar time they were absent.
- Mid-build scope changes. "While we're at it, can we also…" is the most expensive sentence in any project. Even small additions cost more than their nominal hours, because they pull engineers off the critical path and ripple into design and QA.
- Indecision on irreversible choices. Domain, CMS, integrations. These are decisions that lock the architecture; revisiting them at week five means part of the build gets thrown away.
Notice what is not on the list: engineering bugs, hosting issues, browser compatibility. Those exist, but they cost days, not weeks.
Week-by-week: a typical Tier 3 build
For grounding, here is what an 8-week build looks like inside our studio. This is a real cadence, not an aspirational one.
- Week 1 — Discovery sprint. Brief, sitemap, content inventory, integration list, locked goals. One workshop, one async review, one signed-off scope document.
- Week 2 — Discovery wrap + design kickoff. Brand audit, design direction, references, type and colour exploration. First moodboard delivered Friday.
- Week 3 — Hi-fi design begins. Homepage and one inner page through to refined direction. Build team starts the project skeleton, deployment pipeline, CMS schema.
- Week 4 — Hi-fi design continues. Remaining templates. Design lock on homepage and template inner page. Build team starts the homepage.
- Week 5 — Design lock on all templates. Build team is mid-way through; content team starts loading copy.
- Week 6 — Feature-complete build on staging. First end-to-end QA pass. Analytics, schema, CMS handover docs.
- Week 7 — Polish, accessibility pass, Core Web Vitals optimisation, client UAT round.
- Week 8 — Launch week. Final content loaded, redirects validated, DNS cut over, post-launch monitoring.
That is a clean project. Add a week if discovery is rushed. Add two if content is late. Add a month if the brand work is still in flight at kickoff.
The post-launch month
Nobody mentions this in the proposal, but the first month after launch is part of the project. Real users surface edge cases that staging never reveals — a form field that confuses Safari, a tracking tag that misfires on mobile, a CMS field nobody had thought to add. We budget two to three weeks of light follow-up work into every engagement.
Treat the first month after launch as a soft window. Plan the press release for week three, not day one. The marketing push is more effective on a site that has been polished by real traffic than on a site that ships with hidden rough edges.
If the timeline in a proposal looks suspiciously short, ask for the week-by-week. Most agencies cannot produce one, because most agencies have not done the planning. The honest ones can — and the timeline they give you will be a number you can hold them to.
That is the project. The rest is just paying attention.
