The situation

A store losing sales every second its pages took to load

The business in this scenario — an Australian home goods eCommerce store running WooCommerce on WordPress — contacted us after noticing a sharp drop in mobile conversion rate over a 3-month period. Revenue was flat despite increasing ad spend. Their Google Ads manager had flagged a Quality Score drop on landing pages. A quick look at GTmetrix confirmed what we suspected: the mobile experience was broken.

Mobile PageSpeed was 29. The largest contentful paint was happening at 6.8 seconds — long past the point most mobile users abandon a page. Desktop performed better at 54, but still failed core thresholds. CLS errors were causing the add-to-cart button to visually jump as the page loaded, producing accidental taps and abandoned sessions.

Core challenges

Four distinct problems compounding each other

Unoptimised image library
3,200 product images were serving full-resolution JPEGs — some over 4MB — with no compression, no modern formats, and no responsive sizing. Images accounted for 78% of total page weight.
Render-blocking resources
14 JavaScript files and 6 CSS files were loading synchronously in the , blocking first paint for over 3 seconds on mobile connections.
No caching layer
Every page visit hit the PHP/MySQL stack from scratch. No object caching, no page caching, and the hosting plan had no built-in CDN.
CLS and layout instability
Web fonts, late-loading ad scripts, and unsized image elements were causing the layout to shift as the page rendered — directly hurting Google's CLS score and user experience.
Our approach

A structured 6-phase SEO programme

We built a phased plan that addressed technical foundations first, then content, then authority — in that order. Skipping the technical audit would have wasted the content investment.

1
Full performance audit
Used GTmetrix, WebPageTest, and Chrome DevTools to profile load behaviour across mobile 4G and desktop connections. Produced a 22-item fix list with expected impact scores for each item.
GTmetrix auditWebPageTest profilingDevTools waterfallPrioritised fix list
2
Image optimisation at scale
Wrote a batch script to convert all 3,200 product images from JPEG/PNG to WebP format with appropriate quality settings. Implemented responsive srcset attributes so mobile devices received appropriately sized images.
WebP conversionBatch processingsrcset implementationLazy loading
3
Critical CSS extraction and JS deferral
Identified and inlined the critical CSS needed for above-the-fold rendering. Moved all non-critical JavaScript to deferred loading. Removed 4 unused plugins that were loading resources on every page load.
Critical CSS inlineJS defer/asyncPlugin auditUnused CSS removal
4
CDN and caching implementation
Migrated static assets to Cloudflare CDN with appropriate cache-control headers. Implemented Redis object caching for WooCommerce and WP Rocket for full-page caching.
Cloudflare CDNRedis object cacheWP RocketCache-Control headers
5
CLS and font optimisation
Added explicit width and height attributes to all image elements. Preloaded web fonts to eliminate FOUT. Moved third-party scripts (chat widget, analytics) to load after user interaction rather than on page load.
CLS fixesFont preloadingImage dimensionsDeferred third-party scripts
6
Post-optimisation testing and documentation
Ran final tests across 12 device/connection combinations. Documented every change made for the client's developer. Set up ongoing monitoring in Google Search Console and a PageSpeed API alert for regressions.
12-device testingDeveloper documentationSearch Console monitoringRegression alerts
What we would measure

How success would be judged

We agree a baseline in the first week and report against it every month. We don't promise figures before we have seen your data.

Core Web Vitals (field data)
LCP, INP and CLS from real Chrome users, not just a lab score.
Mobile Lighthouse score
Measured on product, category and checkout pages.
Plugin and script weight
Every plugin and third-party script justified or removed.
Conversion rate
Compared before and after, so speed work is judged by sales, not scores.