Ooi Solutions · Case file
A Malaysia-based brand design agency whose website did not reflect the quality of its work. We spent 6 months rebuilding the site and the lead engine behind it.
Baseline: Lighthouse mobile, homepage. Today: PageSpeed Insights field data, 28-day window, Jul 2025.
Source: Maro growth report - GA4, Google Search Console and lead tracking, Feb 2025 → Jul 2025.
Admission notes
Maro designs brands for clients across Malaysia. The irony was that its own website was slow, generic, and invisible to the local searches that mattered most.
Scope
Documented in full in the Maro growth report.
Before & after · the pages
These are the live Maro pages as they were, and as we rebuilt them. Screenshots come straight from the client performance report.
A portfolio-first layout with no clear path to enquiry.
Thin content with no persuasion or social proof.
A single-field form that filtered nobody.
Screenshots: Maro growth report.
Core Web Vitals
Google's threshold for a good LCP is 2.5 seconds. The Maro homepage was more than double it.
Source: PageSpeed Insights baseline, Feb 2025.
Source: PageSpeed Insights, Jul 2025 - real-user field data.
The 6.4s baseline is a Lighthouse lab test of the homepage. The 1.2s is field data. A like-for-like lab re-test would close the loop but field data reflects what real users experience.
Portfolio gallery pages still load heavy image sets. Server-side rendering for project pages would cut LCP further. Breadcrumb schema pending on inner pages.
The fix list
Not recommendations - the actual changes specified, page by page, with the homepage as the reference implementation.
| Issue found | What we changed |
|---|---|
| LCP at 6.4s | Preloaded hero image, converted background to <img>, inlined critical CSS |
| Render-blocking CSS | Extracted critical CSS, deferred non-critical stylesheets |
| Unoptimised images | Converted to WebP, added responsive srcset and sizes |
| No lazy loading | Added loading="lazy" to all below-fold images |
| CLS at 0.32 | Set explicit width and height on all images and embeds |
| Thin service pages | Expanded to 800-1,200 words with client-outcome framing |
| No lead capture | Built multi-step inquiry form with service and budget selection |
| Weak local SEO | GBP overhaul, NAP consistency, LocalBusiness schema |
| No meta strategy | Rewrote title tags and meta descriptions for all pages |
| Missing alt text | Added descriptive alt text to every portfolio and service image |
The unglamorous part
Most of this engagement was infrastructure nobody sees. It is also the reason the rest worked.
Client work uploaded as 4MB TIFFs. Built an automated pipeline to convert, resize and compress to WebP on upload.
Different pages described the same service in different ways. Created a unified messaging framework applied site-wide.
Server was in Europe. Migrated to a Malaysia-based CDN node, cutting TTFB by 340ms.
GA4 was installed but no events were configured. Built a full event taxonomy from scratch.
Three competitor agencies dominated the same SERPs. Differentiated through long-tail service + location combinations.
Portfolio gallery overflows on screens under 400px. Fixed with CSS grid and swipe-friendly layouts.
Results
Every chart below is lifted from the client performance report, drawn from GA4, Google Search Console and lead tracking.
Feb 2025 → Jul 2025
Monthly, Feb 2025 → Jul 2025
Feb-Jul, 2024 vs 2025
Evidence
Search visibility and audience growth, straight from the consoles.
Clicks and impressions trending upward through Jul 2025
Google Analytics · monthly
New vs returning visitors
Ooi Solutions
6 months, 14 interventions, +134% leads - and every number on this page can be checked in GA4, Search Console or PageSpeed Insights. That is the point.
ooisolutions.in · hello@ooisolutions.in