- Market
- Saudi Arabia
- Written for
- Head of Digital, Commercial Director
- Languages
- EN · AR
The symptoms
Four ways a slow website announces itself before anyone measures it
By the time slowness is discussed in a meeting, it has usually been costing money for months. These are the forms the cost takes.
- 01
Paid clicks that bounce into nothing
The campaign pays for the click, the visitor lands on a page still assembling itself, and leaves before it exists. The ad platform reports the click as delivered; the analytics report a bounce; nobody connects the two to the loading time.
- 02
Forms abandoned halfway through
The visitor taps a button and nothing happens, so they tap again, submit twice, or give up. A form that responds a beat late does not read as slow to its owner; it reads as broken to its user.
- 03
A mobile experience the office never sees
Most of your traffic is phones on mobile networks; most of your internal viewing is laptops on office broadband with the site already cached. The audience and the organisation are looking at two different websites, and only one of them pays.
- 04
A team that routes around its own site
Sales starts attaching PDFs instead of sending links; staff apologise for the site in meetings. When the people who own a channel stop trusting it, the channel is already losing quietly, whatever the dashboard says.
Lab and field
The site is fast on your machine because your machine is not the audience
Two kinds of measurement exist, and confusing them is where most slow-site arguments start. Lab data is one synthetic load run by a tool on a machine somewhere — useful for debugging, and unrepresentative by design. Field data is the timings collected from your actual visitors’ browsers as they use the site: their devices, their networks, their distance from your servers. When the two disagree — and they usually do — the field is the one describing the experience your revenue depends on.
For an audience in Saudi Arabia and the Gulf, one field factor deserves naming: distance. A site hosted in Europe or North America with no delivery network makes every visitor in Riyadh pay a geography tax on every request — a floor under the loading time that no amount of front-end tuning lowers. Add the usual weight on top — images exported at capture size, script bundles that block the first tap, custom fonts loaded four ways, third-party tags accumulated one campaign at a time — and the site that feels instant in the office arrives, for its real audience, in instalments.
Reading it yourselves
The evidence you already hold about your own slowness, read in order
Before commissioning anyone, an afternoon with data you already have will tell you whether the complaint is real, where it concentrates, and what class of cause to suspect.
Pull the field numbers
Search Console’s experience report and public field datasets show what real visits to your site measured — no instrumentation required, and no opinion involved. The complaint confirmed or retired with evidence.
Split it by page and device
Sitewide averages hide everything. Separate mobile from desktop and look at the handful of pages that carry enquiries and sales — slowness only matters where the money moves. The problem located, not just confirmed.
Weigh what those pages ship
Any browser’s network panel shows a page’s cargo by class: how much is images, how much script, how much fonts, how much third-party tags. The heaviest class is usually the first suspect. A shortlist of suspects, in order of weight.
Check where it is served from
Find where the site is hosted and whether a delivery network stands in front of it. If the origin is a continent away and nothing caches nearby, part of your slowness is geography, and it is fixed at the delivery layer, not in the code. The geography question answered before the engineering one.
Our role
Where the self-diagnosis ends and a repair engagement properly begins
- 01
The diagnosis, made rigorous
We take the afternoon’s reading and make it decision-grade: full field segmentation, every commercially important page weighed, each delay traced to a named cause rather than a suspected class.
- 02
The honest ranking
Causes are ranked by what fixing each is likely to return on the pages that earn — including the finding nobody sells, which is that some slowness is not worth fixing this quarter.
- 03
The handover to the repair
When the diagnosis says repair, that is a separate engagement with its own scope — our web performance optimisation service — and the diagnosis is written so any competent team, ours or another, could execute it.
Whether the repair happens, and who does it, stays your decision; the diagnosis exists so that decision is made on named causes and expected returns instead of on a feeling everyone shares and nobody can locate.
Common questions
What buyers ask
Everyone here agrees the site feels slow, but every test we run says it is fine. Who is right?+
Probably both, which is the trap. The tests are usually lab runs — one load, from a test machine, often on a fast connection near your hosting — and they honestly report that version of the experience. Your colleagues and customers are reporting a different version: mid-range phones, mobile networks, real distance from the servers, and the site under its full weight of tags and scripts. The tiebreaker is field data, the timings collected from actual visits, which neither the tool nor the frustration can argue with. If the field numbers are healthy on the pages that matter, the complaint may be anecdote; in our experience it more often is not.
Which pages should we look at first when the whole site feels slow?+
The short list where slowness costs money: the pages paid campaigns land on, the enquiry or checkout path, and the two or three pages that carry most organic entries. A site is never uniformly slow — each page has its own weight, its own scripts and its own behaviour — and a sitewide average mixes a heavy campaign page with a light text page until neither is visible. Diagnosing the commercial short list first also keeps the eventual repair honest: work is ranked by what it returns, and a slow page nobody transacts from can wait its turn.
Is the slowness more likely our hosting, our build, or our content?+
Usually a mix, in proportions only measurement reveals — which is why guessing between them wastes budget. Content weight, especially imagery exported at full size, is the most common single offender and the cheapest to fix. Build problems — script bundles that block interaction, fonts loaded badly — show up as a page that looks ready but ignores taps. Hosting and delivery set the floor: a far-away origin with no caching taxes every visit before your code runs at all, which matters doubly for a Gulf audience served from another continent. The diagnosis exists to assign the proportions, because the repair for each lives in a different place and a different budget line.
In this cluster
A diagnosis that replaces "the site is slow" with named causes on named pages, ranked by weight — before any repair budget is committed.
Ask for the diagnosis
