Our own website · Implemented

other types: a clearer digital proposition

An actual renewal of our own website: two service paths, separate languages and a more direct enquiry journey.

Our website centred on Korea–Europe expansion. Website refresh needed a clear place within the same company, without losing the operational business.

Current friction

Geography competed with services in navigation. English and Korean shared content blocks. Two separate contact forms created inconsistent journeys.

What we kept

The company identity, actual offices, registered entity, existing operational scope, SvelteKit and Firebase hosting. Existing articles and their source IDs remain intact.

What we changed

A homepage presenting both service areas, dedicated service-group pages, honest digital work examples, a shared enquiry flow and route-based English, Korean and French localisation.

Technical approach

Prerendered commercial content, server-only CMS retrieval, reciprocal language links, canonical URLs, single-hop redirects, accurate structured data and consent-based analytics.

VISUAL DEMONSTRATION

Before. Findings. After.

Recorded screenshots of the local working versions before and after renewal. They reflect the capture date shown below.

01 / BeforeLocal working version
other types: a clearer digital proposition — before, desktop
02 / AfterLocal working version
other types: a clearer digital proposition — after, desktop

Renewed interface captured in English: 2026-10-08T07:28:47.798Z

ACTUAL LAB MEASUREMENTS

Measure the interface. Observe the business outcome.

Mobile Lighthouse, median of 3 runs
MetricBefore → After
Performance71 → 91
Accessibility93 → 100
SEO100 → 100

Measured on local previews with the same Chrome installation, Lighthouse version and simulated mobile settings. Dates: 2026-10-07T10:48:49.988Z / 2026-10-08T07:28:52.845Z

Measured page: http://127.0.0.1:4173/ (EN)
UI source revision: a5dcc2303bc57d20

Lighthouse 13.5.0 · Chrome/154.0.8037.98 · 412 × 823 · CPU slowdown 4×; simulated RTT 150 ms; throughput 1,638 Kbps; gzip text responses. Fresh browser storage, no analytics choice recorded.

Lab scores do not establish conversion uplift, real-user Core Web Vitals or AI-answer placement. Those outcomes require post-launch observation.