The ads are good: clicks come in, the budget is spent, yet there are fewer enquiries and sales than there should be. Before you change the ads or bids, check one simple thing: how long the page takes to load for a person who lands on it after a click. You pay for every click, and a slow site loses part of those people before they see anything.
Below: what the studies say (and why not to trust them blindly), how to check speed in ten minutes and what you can fix yourself.
What the numbers say
We checked the two studies that are quoted most often against their original sources. Both point the same way, but with caveats.
Portent, 2022. The agency Portent analysed about 100 million page views across 20 sites. The conclusion for online stores: a site that loads in 1 second has a conversion rate (the share of visitors who bought or left an enquiry) 2.5 times higher than a site that loads in 5 seconds. For lead-generation sites the difference is 3 times. This is the same study that Ukrainian blogs retell. The caveat: of the 20 sites only six were online stores, the rest were B2B. Retellings give different conversion figures at five seconds, so we could not find an exact value for 5 seconds in the study itself. Go by the direction, not by the multiplier.
Deloitte for Google, 2020. The study “Milliseconds Make Millions” measured what a gain of 0.1 second (one tenth) gives on a mobile site. For retail sites conversion rose by 8.4%, and shoppers spent 9.2% more. The sample: 37 European and American brands and over 30 million sessions. The caveat: the data was collected at the end of 2019, so it is already several years old. There is also a condition: the 0.1-second gain had to appear on every page of the shopper’s journey across four technical measurements. These were large brands, so for a small store the effect may differ.
We have not tested this on our clients’ accounts, so we present it as other people’s data. There are also older figures like “every second costs this many percent of conversion”. We do not use them: they are about 10 years old and the original source is not visible.
The practical conclusion is simple: speed affects sales, but how many percent depends on your site. So check your own. It takes ten minutes and costs nothing, and it is worth doing before you increase your ad budget.
Step 1. PageSpeed Insights
Open pagespeed.web.dev, paste the address of exactly the page your ad leads to (not necessarily the home page) and run the analysis. Look at the result for mobile devices: most people come from a phone, and the phone test is stricter. The lab tests simulate an average phone on a mobile network.
The report has two sections. “Real-user experience data” comes from actual visitors over the last 28 days. Small sites often do not have it because traffic is low. Below it is “Lab diagnostics”: a simulation that points to what exactly slows the page down.
What the scores mean:
- Google treats a score of 90 and above as good, 50 to 89 as needing improvement and below 50 as poor. Do not chase a perfect 100: the score varies slightly from run to run.
- Three metrics decide whether a page passes the Core Web Vitals check (a set of site usability metrics). LCP: when the largest block appears, usually the main photo or headline; good is up to 2.5 seconds. INP: how quickly the site responds to a tap; good is up to 200 milliseconds. CLS: whether the page “jumps” while loading; good is no more than 0.1.
Fix everything marked red first. Yellow comes after. Do not repair green for the sake of green.
Step 2. Search Console
If your site is verified in Google Search Console (a free Google service that shows how your site appears in search), it has a Core Web Vitals report. It uses data from real visits, separately for phones and computers. Groups of pages get the status of the worst of the three metrics. So one slow image used in all product cards can mark a whole group red.
There are limits. If traffic is low, the report will show “no data available”: that is normal, and then you rely on PageSpeed Insights. Also, after you make fixes the report changes with a delay, because the data is collected over several weeks. Do not panic if nothing changed the next day.
The difference between the two tools is easy to remember. PageSpeed Insights checks one page “right now”, while Search Console shows how it was for real people. Look at Search Console first to find the bad groups, and fix through PageSpeed Insights.
Step 3. What you can fix yourself
Practitioners advise moving from the cheap to the expensive. The first two items are usually within reach of the owner or the site administrator without a developer.
Large images. Photos taken on a phone can weigh several megabytes while being shown small on the page. Compress them with a free online compressor or in your site builder’s settings and choose modern formats (WebP). Photos below the first screen should load only when the visitor scrolls to them: this option (called lazy loading) exists in almost every site management system. Videos are better hosted on YouTube and embedded from there than uploaded to your own hosting.
Unneeded plugins and scripts. Open the plugin list in your site’s admin area and switch off those you do not use. The same goes for widgets: an old chat, counters left over from past ad campaigns, widgets someone once added “for a week”. If your site uses Google Tag Manager (a control panel for tags), open it and check whether it holds tags nobody uses any more. We have a separate article on what should stay connected: what should be connected to your website.
After each change, run PageSpeed Insights again and see what moved. That way you will see what exactly helped.
What you cannot fix yourself. If PageSpeed Insights marks the server response red (in the report this is TTFB, time to first byte: good is up to 0.8 seconds), the problem lies in the hosting or in the code itself. This is where you really need a developer or hosting support. But go to them with the number from the report, not with the words “the site is slow”.
Why this matters for advertising in particular
In advertising the cost of the click is already charged while the person is still waiting for the page to open. If they did not wait, you paid for a visitor who never arrived. There is also a link to Google Ads: among its tips on the landing page (the page an ad leads to), Google’s help lists fast loading. So it is about money and also about how Google rates the quality of your ads.
Hence the order of actions: first make sure the ad page opens quickly, and only then add budget. The other checks for the ad page are collected in the article website checklist before advertising.
Summary
- Open pagespeed.web.dev, paste the address of the ad page and look at the result for mobile.
- Fix everything red first.
- In Search Console open the Core Web Vitals report and see which groups of pages are marked poor.
- Compress large images, switch off unneeded plugins and old scripts.
- Run the check again and compare.
You can read how we set up advertising in the search advertising and Performance Max sections.
Want a free audit of your Google Ads account? Send us a request — we will look at your account and the page your ads lead to and tell you what is worth fixing first.