Why the site feels slow on a phone on mobile data
You open the site you built. It appears instantly. It has always appeared instantly. You have never once experienced this website as slow.
Your customer is in traffic on Third Mainland Bridge, on a phone with 2GB left in the bundle until payday, on a network that is technically 4G and behaving like something considerably worse. She taps the link a friend sent.
Eleven seconds of white screen. She goes back and taps the next result.
You will never see this happen. It will never show up in a complaint, because nobody complains about a website they never saw. The business simply gets fewer enquiries than it should, and nobody ever finds out why.
This module is about closing that gap. It starts by explaining what slow actually is, because "make it faster" is not a task until you know what you are making faster.
Words in this lesson
| Word | What it means, in plain English |
|---|---|
| Payload | The total bytes a visitor downloads to see the page. |
| Request | One trip to the server for one thing — an image, a font, a script. |
| Render | The moment the browser actually draws something. |
| LCP | How long until the biggest thing on screen appears. |
| INP | How long the page takes to respond when somebody taps. |
| CLS | How much the page jumps about while loading. |
| Throttling | Deliberately slowing your connection to simulate a real one. |
Three things make a page slow
1. How many bytes it sends. Every image, font, stylesheet and script is downloaded. A 6 MB page on a slow connection takes a long time, and there is no trick that avoids physics.
2. How many separate requests it makes. Each item is its own trip. Fifty small files can be slower than a few larger ones, because the trips themselves cost time — and on a high-latency mobile connection, the trip costs more than the cargo.
3. When things happen. A page can be light and still feel slow if the important part is drawn last, or if a script blocks everything until it finishes.
Most WordPress sites you meet fail on the first one, and the cause is almost always images. That is Lesson 2, and it is the biggest single win available to you.
The three numbers that describe it
Google publishes three measurements called Core Web Vitals. They are worth knowing because they describe the experience in the way a visitor actually feels it.
Verified against Google's own documentation in September 2026:
| Metric | What it measures | Good |
|---|---|---|
| LCP — Largest Contentful Paint | How long until the biggest thing appears | 2.5 seconds or less |
| INP — Interaction to Next Paint | How long the page takes to respond to a tap | 200 milliseconds or less |
| CLS — Cumulative Layout Shift | How much the page jumps while loading | 0.1 or less |
One thing to know, because it will save you from old advice: INP replaced an older metric called FID in 2024. If a tutorial or plugin talks about First Input Delay, it is out of date, and you should be suspicious of the rest of its advice too.
Check the current thresholds yourself at web.dev/articles/vitals before quoting numbers to a client. These move, and this lesson is a dated snapshot — the same discipline as everywhere else in this course.
CLS is the one you have felt
The other two are self-explanatory. CLS deserves a sentence, because everyone has experienced it without knowing the name.
You are reading a page. You go to tap a link. At that exact moment an image finishes loading above it, everything shifts down, and you tap something else entirely. Sometimes an advert. Sometimes "delete".
That is layout shift, and on a slow connection it happens constantly, because things arrive at different times. The commonest cause on a WordPress site is images without their dimensions set, so the browser does not reserve space for them until they arrive.
It is worth caring about because it is the most annoying kind of slow. A page that takes four seconds is tolerable. A page that moves under your thumb makes people leave.
Why this is a money argument, not a vanity one
Two reasons, and the second is specific to your market.
People leave. A visitor who waits and gets nothing goes to the next result. They do not tell the business. The enquiry simply never happens, and it is invisible in every report.
Your visitors are paying for those bytes. This is the part most international advice never mentions. A customer on a metered data bundle is spending their own money to load your client's website. A 6 MB page is not just slow — it is expensive, for somebody who is counting.
That reframes the whole subject. Every unnecessary megabyte is money taken from the person you are trying to sell to. Put that way, a client who was unmoved by "performance" usually starts listening.
Where to measure
Two kinds of tool, and you need both.
A lab test loads the page in a controlled environment and gives you a score and a list of problems. Google's PageSpeed Insights is the common one. It is useful for the list of what to fix, and honest about what it is: a simulation.
A real device on a real network is the only thing that tells you the truth. Lesson 5 is entirely about this, and it is the lesson that changes how people build.
A score is not the goal. A customer who does not leave is the goal, and it is possible to have a mediocre score and a fast-feeling site, or an excellent score and a page that jumps under the thumb.
What we are actually going to do
The rest of this module, in order of how much it wins you:
| Lesson | What it fixes | Typical win |
|---|---|---|
| 2 — Images | Payload | Very large. Usually most of the problem |
| 3 — Caching | Server work per visit | Large, and cheap to do |
| 4 — Being findable | Not speed — whether anybody arrives at all | Different problem, same module |
| 5 — Testing properly | Finding what is actually wrong | The one that keeps it honest |
Do them in that order. People reach for caching plugins first because they feel technical. Images are usually 80% of the problem and a caching plugin will not touch them.
Common mistakes
| Mistake | What it costs |
|---|---|
| Judging speed on your own laptop and wifi | You never once experience the site as customers do |
| Chasing a score instead of an experience | A good number and a page that still jumps under the thumb |
| Installing a caching plugin before fixing images | The biggest cause is untouched |
| Following advice that mentions FID | It predates 2024; treat the rest of its advice with suspicion |
| Treating speed as a technical nicety | It is lost enquiries, and your visitor's own data bundle |
| Assuming nobody complained means nobody left | People who leave never complain |
The short version
- You will never experience the site the way your customer does. That gap is the whole problem.
- Slow comes from bytes, number of requests, and when things happen.
- Three numbers describe it: LCP ≤ 2.5s, INP ≤ 200ms, CLS ≤ 0.1.
- INP replaced FID in 2024. Advice mentioning FID is out of date.
- CLS is the page moving under your thumb, usually from images with no dimensions.
- Speed is money twice over: lost enquiries, and your visitor's own data bundle.
- Use a lab test for the list of problems and a real phone for the truth.
- Fix images first. Caching plugins feel technical and will not touch the biggest cause.
Try it yourself
Part one — measure what exists. Take the three local business websites you have been studying since Module 1. Run each through a lab test such as PageSpeed Insights, on the mobile setting, and record:
| Site A | Site B | Site C | |
|---|---|---|---|
| Total page size | |||
| LCP | |||
| CLS | |||
| Biggest single item on the page |
Part two — the data cost. Convert each site's total page size into something a client understands. If a visitor's bundle is 2 GB, how many times could they open that home page before it is gone?
Write the answer as one sentence you would actually say to a business owner. "Every person who opens your website spends about 1% of a 500 MB bundle on it" lands very differently from "your page is 5 megabytes".
Part three — feel it. Open one of those sites on your phone with wifi off, standing somewhere with poor signal. Count the seconds out loud until you can read something. Then do the same with your own practice site.
Write down the two numbers. That gap, or the lack of it, is what the next four lessons are about.
You'll know you've got it when you can tell a business owner what their website costs a customer to open, in their own data, and watch them care about speed for the first time.
More free previews
- ▶What a CMS actually is, and why your client needs one15 min
- ▶Domains, hosting and DNS, explained once and properly15 min
- ▶Turn the creative brief into a page list15 min
- ▶Choosing a theme you won't have to fight15 min
- ▶A contact form whose messages genuinely arrive15 min
- ▶Why abandoned WordPress sites get taken over — and how it happens15 min
- ▶Scope and price this specific site before you touch a keyboard15 min