Direct answer: the website development process at WebStackRank runs in 7 phases — discovery, SOW commitment, design, build, QA, launch, post-launch — with calendar dates committed in week one and Friday demos every week. Total timeline ranges 3-14 weeks depending on scope. Every phase has named deliverables and explicit gates between them. This is the actual SOP, not a generic Agile diagram.
Phase 1 — Discovery (1-2 weeks for projects above $30K; 2-3 days for smaller)
What happens: stakeholder interviews, current-stack audit if migration, route map, data model sketch, third-party integration list, hosting plan. Output: a written scope of work (SOW) the client signs.
Common pitfall: skipping discovery to save time. Sites built without proper discovery fail in week 6 when the missing requirement surfaces. The discovery investment pays back in fewer change orders.
Phase 2 — SOW commitment + 50% deposit
The SOW is the contract. Named features, named team, calendar dates, USD price, change-order pricing structure, IP transfer terms. Signed by both sides; 50% deposit clears before design starts.
Common pitfall: signing a SOW with "additional features as needed" language. That is the loophole agencies use to bill more. Insist on explicit scope.
Phase 3 — Design (5-10 days)
Figma at component level — design tokens → primitives → composed routes. Client reviews at component granularity, not page-level mockups. Output: signed-off Figma file, accessibility annotations, responsive variants.
Common pitfall: designing pages without designing the underlying system. Pages designed in isolation create inconsistency that costs more to fix in build than to prevent in design.
Phase 4 — Foundation (3-5 days, parallel with end of design)
Repo, CI/CD pipeline, environment scaffolding, layout primitives, font + image pipeline, schema components, observability hooks. The "boring" infrastructure done first so it stays out of the way for the rest of the build.
Common pitfall: deferring infra. Sites built without proper infra accumulate technical debt visible on day 1 in production.
Phase 5 — Build (3-6 weeks for typical scope)
Routes built in priority order. Friday demos every week. Lighthouse budget enforced in CI so performance does not slide as features ship. Issues tracked in Linear (or whatever client uses); PR reviewed by senior on team before merge.
Common pitfall: feature creep without change orders. Every "small addition" mid-build either changes the scope (and the price + timeline) or is deferred. We sign written change orders before any out-of-scope work begins.
Phase 6 — QA + content load (5-10 days)
Cross-browser (Chrome, Safari, Firefox, Edge). Mobile testing on real devices at 320px, 375px, 414px, 768px, 1024px, 1280px, 1920px. Accessibility pass (axe-core + manual screen-reader). Schema validation in Google Rich Results Test. hreflang verification if bilingual. Real content loaded into the CMS by the client (or by us for an additional scope item).
Common pitfall: skipping accessibility "until later." There is no later. Litigation exposure is real (especially in California and New York). AA baseline is non-negotiable.
Phase 7 — Launch + post-launch (2-3 days + 30 days)
DNS cut-over, GA4 + Search Console wiring, 301 redirect map if migration, repo and credentials transferred. 30-minute Loom hand-off recording. 30 days of bug fixes included.
Common pitfall: deploying on a Friday afternoon. Never. We launch Tuesday or Wednesday morning so the team is available all week for fixes.
What "done" actually means contractually
The SOW lists the delivery criteria:
- Live URL on the client's domain
- Lighthouse mobile Performance ≥ 90 measured on the live URL
- Every page from the SOW visible and working
- CMS-editable where specified
- GA4 firing, Search Console submitted, sitemap valid
- Contact form submitting and emailing to client inbox
- Every form-validation case tested
- Accessibility AA verified
- Schema validation passed in Google Rich Results Test
- 30-minute hand-off recording delivered
If any item fails, the second invoice does not issue. We keep working until it does.
Things that go wrong (honest section)
The two biggest project-failure modes we have seen in 12+ years of shipping:
-
Client-side content delay. Site is ready; copy is not. We have shipped pages with "lorem ipsum" placeholder content because the client never delivered final copy. Solution: build content delivery into the SOW with calendar dates.
-
Scope creep without change orders. Small additions accumulate. We have rebuilt projects where the original SOW would have been 6 weeks and the actual delivery took 14 weeks because of unmanaged scope creep. Solution: written change order before any out-of-scope work.
If this process maps to what you need, the core web development services page is the engagement entry. Or use the quote calculator for an instant estimate.