Website Design & Rebuilds, Engineered for Search
Most websites are built by people who never have to make them rank. Design is delivered, signed off, launched — and some months later a different team is asked why nobody can find it. This page is about what changes when those two jobs sit with the same people.
Splitting the build from search costs you months later
A redesign throws away years of accumulated signal
Old URLs left unmapped with no 301s point years of link and ranking signal at addresses that no longer exist. It's the most common thing we're called in to fix, and it happens after a launch everyone considered a success.
URL and content structure are decided on build day
How pages are split, how keywords map to them, how languages are separated — these feel like incidental choices during design, and changing them after launch means running a migration all over again. The later it's addressed, the more it costs.
Speed is a consequence of design decisions
How many font weights load, what format images are in, how many third-party scripts run — all settled at build time. Chasing Core Web Vitals afterwards leaves far less room than people expect.
Structured data usually belongs to nobody
The design side considers it out of scope; the marketing side can't touch the code. So a site runs for a year with no schema at all — one of the main ways search engines and AI understand what you actually do.
Decided on build day
Every item below can be changed after launch, at several times the cost and with visibility lost in between. So they're handled during the build:
- URL structure and how pages divide the work, planned by search intent so pages don't compete with each other
- For a rebuild, the URL map and 301s that carry existing signal to the new pages
- Multilingual structure and reciprocal hreflang, so language versions don't compete
- Structured data — Organization, Service, FAQPage, BreadcrumbList and the rest
- A speed budget: font weights, image formats and dimensions, how many third-party scripts
- Crawlability: robots rules, sitemap, canonicals, and protection settings that don't block crawlers
- A content model your non-technical colleagues can actually update without calling a developer
We don't recommend a platform
We've delivered on WordPress, Shopify, Wix and custom builds in Next.js. Which one is right depends on what your team can maintain, not on which we happen to prefer.
In most cases the answer is to stay where you are and fix it. Changing platform doesn't solve a tangled data model or an absent content strategy — it moves the same problems somewhere new and adds migration risk on top.
A rebuild is genuinely warranted when plugins have accumulated past the point of untangling, when the data structure can't support the content you need, or when the performance problem is architectural rather than configuration. We'll say so when that's the case — and say the opposite when it isn't.
Both directions, evidenced
What it looks like done right: Amgen Aluminium Windows had been running for over a decade. We rebuilt it and then ran SEO / AEO / GEO; monthly organic clicks went from 72 in March 2026 to 2,021 in July (Google Search Console — we hold access to the property, published with the client's approval). Worth saying that demand for window work is seasonal and July is a traditional peak, so that portion isn't ours alone.
What it looks like done wrong: ID Cosmetic Clinic's redesign had already shipped before we were involved. The design was fine; the old URLs were left unmapped with no 301s, and hreflang was never updated after the language structure moved into subdirectories. We took it over in May 2026. We publish no traffic figures for that one — we don't hold Search Console access to the property, so nothing we quoted could be verified by us.
A site running over a decade, rebuilt and then run on SEO / AEO / GEO. With first-party Search Console data.
Read the caseTraffic lost after a redesign: old URLs unmapped, hreflang wrong. We took it over and repaired it.
Read the caseDelivery and cost
- Standalone design or rebuild: from HKD 8,800 one-off, quoted to scope
- Clients on a 6-month or longer SEO plan: no cost for the build, conditional on completing that plan
- Typically 2–4 weeks, depending on page count and how ready the content is
- Payment: 50% on signing, 50% on launch acceptance
- Website work is billed separately from SEO plans — the two aren't bundled
A standalone build starts at HKD 8,800. It costs nothing for clients who join a 6-month SEO plan and complete it — the condition is part of the offer, not a footnote.
Frequently Asked Questions
Do I have to do both SEO and a website?
Which languages do the websites support?
Should we rebuild, or fix what we have?
Will a rebuild affect our current rankings?
Can we update the content ourselves afterwards?
Who owns the site, the domain, and the data?
Find out whether you need a rebuild at all
The free search-visibility review includes an initial technical and content diagnosis. If the conclusion is that optimising what you have is enough, we'll say so — there's no reason to recommend a rebuild that doesn't pay for itself.


