Why a Slow Website Costs You Money and What Actually Affects Speed
A slow site loses customers before they have even seen it. A study by Google and Deloitte found that even a one tenth of a second improvement moves the business numbers up. The thing that slows sites down most is almost always the same one: images that are too heavy. It is also the easiest thing to fix.
What Google actually measures
Google does not measure speed as a single number but as three metrics it calls Core Web Vitals. Each answers a different question the visitor asks without ever putting it into words.
It is worth knowing that one of the metrics changed. Until 2024 the second one was FID and since then it is INP. If you come across a guide that talks about FID, it is out of date.
| The metric | The threshold for a good page | What it answers |
|---|---|---|
| LCP | Up to 2.5 seconds | How long until the main content of the page loads. In practice the big image or the headline |
| INP | Up to 200 milliseconds | How long the page takes to respond after a tap. Replaced FID in 2024 |
| CLS | Up to 0.1 | How much the content jumps while loading. When you go to tap a button and it moves at the last second |
The detail everyone misses about the measurement
Google does not test your load time from your office. It measures the **75th percentile of the real loads** of your actual visitors, separately for mobile and desktop.
The meaning matters. It means a quarter of your visitors can be having a worse experience than the number shows. And it means one test from your own computer on a fibre connection tells you nothing. Your visitor is standing in the street in Eilat with average reception on a three year old phone.
A new site usually has no such field data at all, because there is not enough traffic. In that case what is left is lab tests that simulate a phone, and you have to remember they are an approximation and not the truth.
What it is worth in money
This is the part usually said without a source, so here is the source. Google together with Deloitte Digital and Fifty-Five analysed four weeks of mobile site data from 37 brands in retail, travel, luxury and lead generation, across Europe and the United States. They isolated speed as the variable and checked what moved with it.
The finding: **an improvement of a single tenth of a second** in load time moved the entire funnel.
- In retail, conversions rose 8.4% and average order value 9.2%
- In travel, conversions rose 10.1%
- In luxury, progression to the contact page rose 20.6% and from product page to basket 40.1%
- And in lead generation, the line relevant to most businesses in Eilat, **the bounce rate on informational pages improved by 8.3%**
Why the lead generation line is the one that matters
Most of the businesses we work with do not sell through a shopping cart. A restaurant, a carpentry workshop, a water sports club or a transfer company want people to call them or fill in a form. In the language of the study that is lead generation, and there the number that moved is the bounce rate on informational pages.
That makes sense. Someone who landed on your service page and waits two unnecessary seconds simply goes back to the results and clicks the next one. They are not angry and they will not remember you badly. They just left, and you will never know about it.
Which is why it matters especially in Eilat. A large share of searches happen when the tourist is already in the city, looking on their phone for something close by right now. We expanded on that moment in our demand map. In a search like that there is no patience at all.
What actually slows sites down
After quite a few sites we have checked, the picture repeats itself. The order here is by how much it really matters and not by how much it gets talked about.
- **Images. This is reason number one by a wide margin.** A photo uploaded straight from the camera weighs several megabytes and is displayed at the size of a business card. A modern format and the right dimensions cut that by tens of percent without touching anything else
- Too many plugins or external scripts. Every tracking tool, pixel and chat widget adds code that runs on the visitor’s device. Three of those is fine. Twelve is a slow site
- Fonts loaded from outside in too many weights. Every weight is a file. Four weights of two fonts is eight files before a single word is written
- Hosting that is too cheap. If the server answers slowly, everything after it is delayed. That shows up as TTFB and no other optimisation can compensate for it
- Video that plays automatically. Even compressed it is the heaviest thing on the page
Our own check on ourselves
The fairest way is to check on a real site, so we measured ours in September 2026. We ran the pages in a browser on a simulated phone with the processor slowed by a factor of four and a slow mobile connection, which brings the experience closer to a mid range device than to a new and fast one.
| What was measured | Home page | An article page | Google’s threshold |
|---|---|---|---|
| LCP on mobile | 1.94 seconds | 1.52 seconds | Up to 2.5 seconds |
| LCP on desktop | 0.56 seconds | 0.43 seconds | Up to 2.5 seconds |
| CLS | 0 | 0 | Up to 0.1 |
| Server response | 316 ms | 248 ms | No official threshold |
What this means and what it does not
The numbers in the table are in the good range and we are pleased with them. But it is worth saying the other side too, so you do not take more from this than there is.
**These are lab measurements and not field data.** Google judges by real loads from real visitors at the 75th percentile, and this site still does not have enough traffic to have such data at all. A test on a simulated phone is a good approximation and it is not the same thing.
And one more thing is worth saying. **Speed will not rank you first.** Google says itself that page experience is one signal among many, and good content beats a fast site with weak content. Speed prevents a loss, it does not create a win.
What to do in practice, in order
If you have a slow site and you want to know where to start, this is the order we work in. The first ones give most of the result.
- Compress the images and convert them to a modern format. This is the step where you see the big jump, and usually one pass is enough
- Make sure the image loads at the size it is displayed at and not at its original size
- Go through the plugins and scripts and remove everything not in use. The tracking tool installed once for a test and left behind
- Cut font weights down to two or three in total
- Check the server response time. If it is high, changing hosting is worth more than all the rest combined
- Only then go into the fine optimisations. They are worth single digit percentages and are not the place to start
And what is not worth the effort
A few pieces of advice come up again and again and they consume time without returning any. Two stand out.
**Chasing a score of 100.** The score is a diagnostic tool and not a target. A site that scored 92 and loads in a second and a half is exactly as good as one that scored 100. The road from 92 to 100 usually goes through giving up things the visitor actually wants.
**Rebuilding the site just for speed.** In most cases we have seen, the existing site is perfectly fine and the problem is the images and the plugins. We expanded on when a rebuild is worth it in our guide on WordPress or custom code. A new site built properly will be fast, but that is not a reason to build it.
This is exactly what we do at Website building in Eilat. If you want us to go over it together for your business, write to us.
Frequently asked questions
Google’s practical rule is that the main content of the page should load within 2.5 seconds. That is measured at the 75th percentile of real loads, meaning a quarter of your visitors will still experience something slower than that.
Images, by a wide margin. A photo uploaded straight from the camera weighs megabytes and is displayed small. Compression and conversion to a modern format give most of the improvement before you touch anything else.
Yes, but less than people tend to think. Page experience is one signal among many, and good content on an average site will beat weak content on a fast one. Speed prevents a loss more than it creates a ranking.
No. The score is a diagnostic tool and not a target. A site with 92 that loads in a second and a half is exactly as good as one with 100, and the last ten points usually cost you things the visitor actually wants.
INP measures how long the page takes to respond after someone interacts with it. It became a stable metric in 2024, replacing FID, which only measured the delay of the very first interaction. If a guide talks about FID, it is out of date.
In most cases no. Usually the problem is heavy images and unnecessary plugins rather than the platform. Building a new site is the right decision for other reasons, and if it is being made anyway then a properly built site will also be fast.
Want us to look at this for your business?
Just write to us on WhatsApp. The first conversation comes with no commitment. 050-294-0127
Write to us on WhatsApp