We measured 14 Egyptian and Gulf WordPress sites every day for a month
Not a survey and not an estimate. Every site below was measured with the same tool, on the same simulated phone, every morning. Here is what is actually making WordPress slow in this region — and it is not what most people are told.
- 67
- Median mobile performance score across the 14 sites. The desktop median is 85.
- 2 of 14
- Sites with a Largest Contentful Paint inside Google's good range of 2.5 seconds.
- 13 of 14
- Sites that pass layout shift. Visual stability is not the problem here.
Finding one
The phone is where the damage is
The median site scores 85 on desktop and 67 on mobile — a gap of 17 and a half points. Owners check their site on a laptop on a good office connection, see it load instantly, and conclude it is fine. Their customers are on a phone on mobile data, and that is the number Google ranks on.
Finding two
Half these sites take longer than four seconds to show their main content
Largest Contentful Paint is the moment the biggest thing on screen finishes loading. Google calls 2.5 seconds good and anything over 4 seconds poor. Two of the fourteen sites are good. Seven are poor. The median is 4.7 seconds, and the worst reading was 12.3 seconds on a site whose first paint was a very fast 1.1 seconds — it drew the page quickly and then sat there, blocked, for eleven more.
Finding three
Hosting is not the bottleneck, and neither is layout shift
Two things people spend money on turned out to be fine. Median time to first byte was 131 milliseconds, which is a healthy server response. And 13 of the 14 sites pass Cumulative Layout Shift. If you are being sold a hosting upgrade to fix your speed score, ask to see the time-to-first-byte number first.
Finding four
It is other people's JavaScript
In every slow site we pinned the cause to code loaded from someone else's server. These are the six worst offenders we measured, and none of them is the website's own code:
| Three Google Tag Manager containers on one site | 1,116 ms blocking · 502 KB |
|---|---|
| A single embedded YouTube video | 1,041,445 bytes |
| One third-party chat and reviews widget | 975 ms render-blocking, on every run |
| Google Fonts loaded from Google's servers | 763 ms render-blocking |
| A hosting provider's own injected script | 933 ms render-blocking |
| A Facebook pixel plus a session recorder | 495 KB combined |
The tag manager case is worth sitting with. Three containers had been added to one site over the years by three different people, and none had been removed. Together they cost that site more than a full second of blocked main thread on every visit.
Finding five
The fastest site in the set makes 19 requests
The median site makes 90 requests. The fastest makes 19, scores 92 on mobile and 100 on desktop, and has a total blocking time of 6 milliseconds. It is the one site in the group that is not WordPress — it is a static site. We are not arguing that everyone should leave WordPress. We are pointing out that almost all of the gap is request count and third-party JavaScript, and both are choices rather than facts.
How this was measured
Method, in full, so you can check it
Lighthouse 12.8.2 command line, pinned to that version. Google's standard mobile profile — a 4× CPU slowdown and a simulated slow 4G connection — plus the desktop preset. Cache warmed twice before each measurement. Between two and four runs per site per day, median taken. One throwaway run discarded at the start of every session, because the first measurement on a cold machine always reads high. Where a site's score moved, we confirmed it against sites serving byte-identical HTML that day, so that a slower machine could never be mistaken for a real regression. Lab data only — none of these sites has enough traffic for Google's field data.
What this is not. Fourteen sites is a small sample, and they are all sites we maintain, which is not a random selection of Egyptian WordPress. Treat the direction as solid and the exact figures as indicative. Sites are unnamed on purpose — several are clients, and the diagnosis is the useful part, not the league table.
Where to go next
Tell us what you want to build.
A website, a mobile app, a CRM or an ERP: tell us what you have in mind and we'll come back with a plan, a timeline and a price. Thirty minutes is enough to tell whether we're the right team.