From 2.2s to 0.13s — full infrastructure overhaul across 6 production sites.
A portfolio of high-traffic sites was bleeding performance and money. We migrated the entire hosting architecture, implemented edge caching, and cut load times by 94%.
The short version.
A portfolio company managing 6 high-traffic WordPress sites was experiencing degraded performance, escalating hosting costs, and poor Core Web Vitals across all properties. Page load times averaged 2.2 seconds, conversion rates were declining, and the existing Apache-based infrastructure couldn't scale.
Following an initial engineering audit to map the performance bottlenecks, we executed a comprehensive infrastructure overhaul: migrated from Apache to OpenLiteSpeed, implemented Cloudflare edge caching, restructured DNS and CDN configuration, and optimized every layer from server to browser. The result: 0.13s average page load, 3× conversion lift, and 40% reduction in monthly hosting costs.
Performance was hurting the business.
Six production sites on shared hosting with Apache. Each site had accumulated years of plugin bloat, unoptimized images, and render-blocking resources. The symptoms were clear:
- 2.2-second average TTFB: Google was penalizing rankings
- Core Web Vitals failing: LCP, FID, and CLS all in red
- Conversion rates dropping: every 100ms of delay was costing revenue
- Hosting costs escalating: scaling Apache horizontally was expensive
- No caching strategy: every request hit the origin server
The client had tried quick fixes: caching plugins, image optimization plugins, CDN bolt-ons. None addressed the root cause: the infrastructure itself was the bottleneck.
Full-stack infrastructure redesign.
We didn't optimize the existing setup. We replaced it. The migration happened in phases to ensure zero downtime across all 6 sites:
- Apache → OpenLiteSpeed: native HTTP/3, built-in page caching, 3× lower memory footprint
- Cloudflare edge caching: full-page cache at 300+ edge locations, automatic purge on content update
- DNS restructuring: CNAME flattening, optimized TTLs, geographic routing
- Asset pipeline: critical CSS inlining, deferred JS loading, WebP conversion, responsive image srcsets
- Database optimization: query profiling, index tuning, transient cleanup, object caching with Redis
- Monitoring: real-time performance dashboards, uptime alerts, Core Web Vitals tracking
Layered caching with edge-first delivery.
The architecture follows a three-tier caching strategy: Cloudflare edge cache handles 95%+ of requests, OpenLiteSpeed's built-in LSCache serves the remainder, and Redis handles object-level caching for dynamic queries.
Three-tier caching | 95% of requests served at the edge
This means the origin server handles less than 5% of total traffic. The result: consistent sub-200ms response times regardless of traffic spikes, and infrastructure costs that scale logarithmically rather than linearly.
"The difference was immediate. Pages that took 2+ seconds now load instantly. Our conversion rates tripled and our hosting bill went down. Synthax didn't just optimize. They rebuilt everything the right way."
Portfolio Manager, details anonymized per NDA
Have a similar
performance problem?
We'll audit your current setup and tell you exactly what's possible, and what it costs.