Website speed and conversions: how much does every second cost you?
Website speed affects conversions most strongly where you have paid for the visitor. With paid traffic every lost second is a wasted lev of ad spend, because the person has already clicked and the fee has been charged. Two thresholds matter for sales: the main content should appear in under 2.5 seconds and the layout should not jump, that is CLS under 0.1. Mobile speed weighs more than desktop, because most paid clicks come from a phone, on a weaker network and a weaker processor. That is why in an audit I look at the speed of the product page and the checkout before the home page. They decide the result, while the home page usually gets the most attention for no reason. The check is simple: compare the conversion of fast and slow sessions.
My name is Hristiyan Panov — a performance marketing consultant with over 5 years’ experience and over €2 000 000 Managed advertising budgets for over 90 companies on Google and Meta. Below, I explain which metrics are meaningful, which are just noise, and where the delays are coming from for most Bulgarian online shops.
Which metrics are significant for sales?
- LCP — how quickly the main content appears. This is what the user perceives as „has it loaded yet?“.
- INP — how quickly the page responds to an action. This is important for filters, options and adding items to the basket.
- CLS — whether the content jumps around. This can lead to accidental clicks and abandoned shopping baskets on mobile devices.
The overall score from speed tools is useful as a guide, but it is not an end in itself. I have seen shops with mediocre scores and excellent conversion rates — and vice versa.
Why is mobile speed more important than desktop speed?
This is because, for most Bulgarian shops, mobile traffic accounts for the majority of their traffic, and advertising on Meta is almost entirely mobile-based. The desktop version usually looks good and creates a false sense of security.
First check the ratio in GA4: a breakdown of sessions by device, done separately for the paid source. If mobile users are the majority there, desktop speed is secondary.
Then test under the real conditions of the click. A click from Meta does not open Chrome, but the app’s built-in browser, which starts more slowly and rarely uses files that have already been downloaded. Open exactly the address that is in the ad, together with the UTM parameters and all the redirects along the way.
The practical rule: test over a mobile network, not office Wi-Fi, and on a mid-range phone, not the latest model. Measure with a cleared cache and on the product page, not on the home page.
Finally, compare the lab measurement with the data from real visitors. A gap between the two usually means you are testing under conditions better than those of the client.
What is the most common cause of delays?
- There are far too many tracking scripts. Pixels, chat sessions, retargeting tags and tests, built up over the years, without anyone ever removing them.
- Unoptimised images. Product images in their original size, resized using CSS.
- Heavy-duty page builders. Every additional layer adds code that the browser has to execute.
- Web hosting that can’t cope with traffic peaks. It becomes apparent precisely when the advert is at its most effective.
- Fonts and animations, which are loaded before the content.
What should you sort out first?
The ranking I use is based on the ratio of effect to effort:
- Removing unused scripts — fastest result, smallest risk. Make a list of everything that loads on the product page, and next to each line write down who uses it today. Lines with no owner go.
- Image compression and modern image formats. Export them at the size at which they are actually displayed, with a separate variant for phones, and set the width and height, so the layout does not jump.
- Server-level caching. The catalogue is cached; the cart and the checkout are excluded, so that another person’s session is not served.
- Putting aside everything that isn’t needed for the first screen. Chats, maps, reviews and animations are loaded after the user’s first action.
- Changing hosting — the most expensive and the least often the root cause.
Work on one thing at a time and measure before and after, so you know what produced what. After each deferral, check that the pixel and GA4 still send the purchase — otherwise you gain seconds and lose data.
How can you check whether your speed is costing you money?
Compare conversion rates by device in your analytics. If the mobile conversion rate is significantly lower than the desktop rate for the same traffic source, the difference is rarely down to people’s intentions — it’s more often down to the user experience.
To make the comparison fair, look at the same period, the same source and without branded traffic — it converts higher and distorts the average.
Then see exactly where they stop: product page, cart or checkout. That narrows the search from “the website is slow” to a specific page.
In GA4 arrange the steps one after another — product view, add to cart, checkout started, purchase — and see at which step the drop-off on phones is bigger than on desktop. That step points to the page to measure.
Finally, connect the two. Take the pages the ads lead to and set the CPA of the corresponding ad set against the time to the main content of the page. If the most expensive results come from the slowest pages, you have the answer and the order of work.
Does page speed help me rank higher on Google?
It is one of the factors, but not the decisive one. Given equally good content, the faster page has the advantage; a faster page with poor content does not outperform a slower one with strong content. With paid traffic, however, the effect is direct and immediate — there, you pay for every visitor who leaves.
FAQ — Frequently Asked Questions about Speed and Conversions
What speed is sufficient?
A good rule of thumb is that the main content should load in under 2.5 seconds on a mobile device with an average network connection. It’s more important to be faster than before than to chase a score.
Do pixels slow down the website?
Yes, especially when there are several of them and they load simultaneously. Moving them to a tag manager and delaying their loading solves a large part of the problem.
Will my ad performance improve if I speed up the website?
Usually, the conversion rate improves whilst traffic remains the same, which reduces the cost per conversion. The extent of this improvement depends on how poor the conversion rate was to begin with.
Is a separate mobile version worth it?
These days, it’s rare. A responsive design that prioritises the mobile experience is usually a better solution.
Who should do it?
A developer for the technical side, but priorities are determined by conversion data, not by the speed tool.
Conclusion
Focusing on advertising without optimising your website is like advertising a shop with a broken window. Speed isn’t an end in itself — it’s one of the few changes that simultaneously improves paid traffic, organic traffic and the customer experience.
If you’re paying for traffic and don’t know where it’s going, start by getting an independent review.
Request a quote for a digital audit →
Related articles: Pixel and CAPI · Consent Mode v2 · advertising account audit
