LiteSpeed vs Apache: what the numbers look like on real Kenyan traffic
We ran the same WordPress site on Apache and LiteSpeed from Nairobi and measured time to first byte, page load and CPU load. Here is what changed, and where caching mattered more than the web server.
The test setup
Same WordPress build, same theme, same plugins. One host running Apache with mod_php, one running LiteSpeed Enterprise with LSCache. Both on NVMe storage in Nairobi, tested from a 20 Mbps fibre line in Westlands and a 4G connection in Kilimani.
What we measured
- Time to first byte (TTFB) — Apache averaged 340 ms on uncached pages, LiteSpeed 155 ms.
- Full page load — 1.8 s versus 0.9 s with the same page weight.
- CPU load under concurrency — LiteSpeed handled roughly 3x the concurrent requests before load average crossed 1.0.
The bigger lesson
Caching mattered more than the web server choice. With LSCache enabled, TTFB dropped to 40–70 ms and the site felt instant on both the wired and mobile connections. The win was not just the server; it was that LiteSpeed ships a caching layer that most WordPress users never switch on.
What to do on your own site
- Turn on the LiteSpeed Cache plugin and enable page caching.
- Serve images as WebP — it typically cuts 40% of page weight.
- Keep PHP to one version and enable OPcache.
- Measure before and after. Feelings are not data.
If your current host cannot give you a caching layer, that is usually more expensive than the difference in licence fees.