Mobile SEO is the work that makes a site rank and perform when the search starts on a smartphone: how Google reads the page, how fast it is on a phone, how easy it is to use with one thumb and how it appears in the mobile SERP.
Part of that work was settled more than two years ago. Since July 2024 Google has indexed every site with its smartphone crawler, and the tool people used to check whether a page was "mobile-friendly" no longer exists. Many guides still spend most of their time on that part, and at least one on page one of Google UK still recommends the retired test. Hardly any look at the other part: the results Google shows on a phone are often different from the ones you see when you check your own site from a computer. On 3 October 2026 we measured how different, across 20 searches in Italy, and on 10 October we ran a smaller check on Google UK.
What's left of mobile SEO after mobile-first indexing
Mobile-first indexing decides which version of the page Google reads: the one served to a smartphone. It's an indexing issue and we cover it fully in our guide to mobile-first indexing, with a test on 49 Italian pages in which no site still used a separate m. version. Here it's worth remembering that three pieces of advice that are still widely quoted have expired:
- The Mobile-Friendly Test, its API and the Mobile Usability report in Search Console were retired on 1 December 2023, as Google announced in April 2023. Google made clear that mobile usability still matters and pointed to Chrome's Lighthouse as one of the tools for assessing it.
- FID is no longer a Core Web Vital: on 12 March 2024 it was replaced by INP (Interaction to Next Paint). Guides that list LCP, FID and CLS describe a metric Search Console no longer reports.
- You don't need AMP to get into the Top Stories carousel: Google wrote in 2021 that the AMP format is no longer required and that any page that meets the Google News content policies can appear.
What remains is easier to list and harder to do: the page has to load quickly on a mid-range phone, be usable without zooming, not cover the content with pop-ups, and be measured with data split by device. Plus one thing that projects rarely check: your ranking on a smartphone and your ranking on a desktop are not the same.
How much the SERP changes between phone and desktop: 20 searches in Italy
The mobile SERP is the results page Google shows to people searching on a smartphone. Its layout differs from the desktop one and, as we found, so does its order, quite often.
We chose 20 Italian searches of three types: 7 local (such as "dentista roma" (dentist rome) and "palestra torino" (gym turin)), 9 commercial or product searches (such as "assicurazione auto" (car insurance) and "scarpe running uomo" (men's running shoes)) and 4 informational ones (such as "730 precompilato" (pre-filled tax return) and "ricetta carbonara" (carbonara recipe)). On 3 October 2026, between 23:13 and 23:22 UTC, we downloaded Google's results page for each one twice with the DataForSEO SERP API, once as desktop and once as smartphone, with Italian as the language and Italy as the location. Then we compared the top 10 organic results URL by URL, ignoring protocol, "www", parameters and trailing slash.
| Indicator | All (20) | Local (7) | Commercial (9) | Informational (4) |
|---|---|---|---|---|
| URLs in common in the top 10 (mean) | 2.6 | 3.4 | 2.9 | 0.5 |
| URLs in common in the top 10 (median) | 1 | 2 | 1 | 0.5 |
| Searches with zero URLs in common | 8 | 2 | 4 | 2 |
| Same first organic result | 4 (+1) | 2 (+1) | 2 | 0 |
| URLs in common in the top 3 (mean) | 0.8 | 1.1 | 0.9 | 0 |
| Local pack shown (desktop / mobile) | 8 / 10 | 6 / 6 | 2 / 4 | 0 / 0 |
| AI Overview shown (desktop / mobile) | 7 / 6 | 0 / 0 | 3 / 3 | 4 / 3 |
In 13 of the 20 searches the two lists share one URL at most. Only three are almost identical, with 9 of 10 URLs shared: "commercialista torino" (accountant turin), "software gestionale magazzino" (warehouse management software) and "noleggio auto lungo termine" (long-term car leasing). Just one, "software gestionale magazzino", also has the same order. At the other end are the informational searches: half a URL in common on average, and no shared first result.
The test has two limits worth stating. It's a single snapshot, taken from a server with no search history, so it isn't personalised: someone searching on their own phone sees an even more specific page. And the two SERPs were downloaded a few minutes apart, so some of the difference may come from the normal churn in results rather than from the device. The three near-identical searches suggest that churn alone can't explain eight lists without a single URL in common.
We ran the same comparison on Google UK on 10 October 2026 for three searches, one of each type, with English as the language and the United Kingdom as the location. "car insurance" had zero URLs in common between desktop and mobile: on desktop the first organic result was Aviva's quote page, on mobile it was Diamond's homepage, and Compare the Market, GoCompare and MoneySuperMarket appeared on both devices but with different pages. "plumber manchester" also had zero URLs in common, even though the local pack showed the same three businesses on both. "self assessment tax return" shared 6 URLs, with the same Which? guide in first place. Three searches prove nothing on their own. Still, the commercial and local searches behaved as they did in Italy, while the informational one, unlike the Italian ones, barely moved.
The 20 searches, one by one
| Search | Type | URLs in common (of 10) | Same first result | Local pack (desktop / mobile) |
|---|---|---|---|---|
| impresa di pulizie milano (cleaning company milan) | Local | 1 | No | Yes / Yes |
| commercialista torino (accountant turin) | Local | 9 | Yes | Yes / Yes |
| dentista roma (dentist rome) | Local | 2 | Same page, different parameter | Yes / Yes |
| idraulico napoli (plumber naples) | Local | 0 | No | Yes / Yes |
| ristorante pesce bologna (fish restaurant bologna) | Local | 0 | No | Yes / Yes |
| hotel firenze centro (hotel florence centre) | Local | 5 | Yes | No / No |
| palestra torino (gym turin) | Local | 7 | No | Yes / Yes |
| serramenti in pvc prezzi (upvc windows prices) | Commercial | 6 | No | No / No |
| macchine caffè professionali (commercial coffee machines) | Commercial | 1 | No | No / Yes |
| scarpe running uomo (men's running shoes) | Commercial | 1 | No | No / Yes |
| condizionatore 12000 btu (12000 btu air conditioner) | Commercial | 0 | No | No / No |
| assicurazione auto (car insurance) | Commercial | 0 | No | Yes / Yes |
| mutuo prima casa (first-time buyer mortgage) | Commercial | 0 | No | No / No |
| software gestionale magazzino (warehouse management software) | Commercial | 9 | Yes | No / No |
| noleggio auto lungo termine (long-term car leasing) | Commercial | 9 | Yes | Yes / Yes |
| corso inglese online (online English course) | Commercial | 0 | No | No / No |
| 730 precompilato (pre-filled tax return) | Informational | 1 | No | No / No |
| ricetta carbonara (carbonara recipe) | Informational | 1 | No | No / No |
| cos'è la seo (what is seo) | Informational | 0 | No | No / No |
| bonus ristrutturazione 2026 (home renovation tax bonus 2026) | Informational | 0 | No | No / No |
Where the two SERPs split
The averages hide cases that, for anyone working in those markets, would change a monthly report. Four examples from the Italian test.
"assicurazione auto" (car insurance). On a smartphone the first organic result is Genertel's product page, followed by those of Prima, Unipol, Conte.it, Linear and Allianz Direct: insurers' pages and quote pages. On desktop none of them appears in the top 10. The first organic result is a special-report page from Corriere della Sera, below a local pack of agencies in Rome, and the list is made up almost entirely of articles and guides, some from 2013 and 2020. We repeated the desktop search a few minutes later and got the same result. Zero URLs in common.
"macchine caffè professionali" (commercial coffee machines). Aranzulla's guide is the first organic result on a computer and the sixth on a phone. The first result on a phone, a page from Espresso Coffee Shop, isn't there at all on a computer. On the phone Google also adds a local pack with four shops that's missing on desktop.
"impresa di pulizie milano" (cleaning company milan). The only URL in common is a Happy Clean page: tenth on desktop, second on mobile. The first desktop result, Aurora Cleaning, only appears on the phone as a listing in the local pack.
"hotel firenze centro" (hotel florence centre). Here the organic results are similar (5 URLs in common), but on desktop Google opens the page with two hotel blocks and the first organic result drops to the third element on the page. On a smartphone the first organic result, Booking.com's page for the area around the Uffizi, is the very first element.
One finding points the opposite way to what you often read: in our sample AI Overviews appear on both devices almost every time (7 searches on desktop, 6 on mobile, with a single difference), and in the UK check they appeared on both devices for "car insurance" and "self assessment tax return". We've covered their effect on clicks in our guide to AI Overviews. What changes most between phone and computer, according to these data, is the list of sites, and on commercial searches the presence of the local pack too: for "scarpe running uomo" (men's running shoes) and "macchine caffè professionali" only people searching on a smartphone see it.
In the UK, four visits in ten still come from a computer
Lots of guides open with a line like "most traffic is mobile". In the UK it's true, but by less than you'd think. According to Statcounter, in September 2026 53.56% of page views in the UK came from smartphones, 41.78% from desktops and 4.66% from tablets. In the United States, in the same month, the split was almost even: 49.1% mobile and 48.47% desktop (Statcounter, United States). In Italy, where we ran the main test, it was 51.78% mobile and 46.61% desktop.
The practical consequence is that a UK site has two audiences of similar weight that see two different rankings. A B2B company that gets enquiries during office hours may well get more organic visits from desktop. For a restaurant or a plumber the opposite is much more likely. To find out which side your site is on you don't need a national statistic: the Performance report in Search Console, filtered by device, is enough. We go through it further down.
Core Web Vitals: the phone is where they fail
Core Web Vitals are the three metrics Google uses to measure how a page loads, responds and stays stable: LCP, INP and CLS. The thresholds for a "good" result, set out by web.dev, are an LCP within 2.5 seconds, an INP of 200 milliseconds or less and a CLS of 0.1 or less, measured at the 75th percentile of visits and separately for mobile and desktop.
The gap between the two devices is clear. According to HTTP Archive's 2025 Web Almanac, which works on real data from the Chrome User Experience Report across millions of sites worldwide, 48% of sites passed all three Core Web Vitals on mobile in 2025, against 56% on desktop. The gap is closing slowly: in 2021 the figures were 32% and 41%.
| Metric | Sites rated "good" on mobile | Sites rated "good" on desktop |
|---|---|---|
| All three Core Web Vitals | 48% | 56% |
| LCP (loading) | 62% | 74% |
| INP (responsiveness) | 77% | 97% |
| CLS (visual stability) | 81% | 72% |
The row to look at is INP: almost every site responds well to clicks on a computer, while nearly one in four doesn't on a phone. A mid-range smartphone's processor takes longer to run the same JavaScript, and the problem only shows up in the mobile data. CLS is the only metric where mobile does better.
How much does all this weigh on rankings? In its page experience documentation Google says that Core Web Vitals "are used by our ranking systems", and that other aspects of page experience "don't directly help your website rank higher". In practice they count when two pages are equal on content and authority, and they count straight away on conversions. In a case we described in our guide to page speed, a mobile LCP alert was closed in ten minutes: the cookie banner covered almost the whole phone screen and Google measured it as the page's main element. On desktop the same banner caused no trouble. The full explanation of the three metrics is in our guide to Core Web Vitals.
What to check on the page, from a phone
Nearly all the mobile SEO problems we find during an SEO audit fit into this table. The thresholds come from the documentation cited, not from house rules.
| Check | Reference | How to verify it |
|---|---|---|
| Meta viewport present and correct | Without a viewport the phone shows the desktop version shrunk down | Page source, Lighthouse |
| Same main content on mobile and desktop | Google asks for the same content on both versions | URL Inspection in Search Console, crawler with a smartphone user agent |
| Text not loaded on interaction | Google doesn't load content that needs a swipe, click or typing to appear | URL Inspection: search the HTML for a sentence from the hidden text |
| Buttons and links big enough | About 48 device-independent pixels according to web.dev; at least 24 x 24 CSS pixels under WCAG 2.2 | Lighthouse, a manual test with your thumb |
| No pop-up covering the page | Google advises against interstitials that hide all the content; cookie consent and age checks are allowed | First visit in incognito mode on a phone |
| LCP, INP and CLS within thresholds on mobile | 2.5 s, 200 ms, 0.1 at the 75th percentile | Core Web Vitals report in Search Console, PageSpeed Insights |
| Images sized for the screen | srcset and sizes, compressed formats | Lighthouse, Network panel in Chrome DevTools |
Button size deserves a closer look, because this is where SEO meets accessibility: a contact form with fields packed too tightly is a problem for anyone using a phone one-handed, and a loss of enquiries that no ranking report will show. For images, the practical advice is in our guide to image optimisation.
Titles, snippets and local results in the mobile SERP
Google truncates result titles "as needed, typically to fit the device width", as it says in its title link guide. There's no guaranteed character count. Our rule is to put the keyword and the information that earns the click in the first few words of the meta title, because on a narrow screen the rest may disappear.
The second point concerns local results. In our Italian test the local pack appeared in 10 of 20 searches on a smartphone and in 8 on desktop, and on local searches the first organic result started lower down on the phone than on the computer in 5 cases out of 7. On Google UK, for "plumber manchester", the mobile page opened with directory links and the local pack before any organic result, while on desktop Google's sponsored Local Services ads came first. For a business with premises, the Google Business Profile listing is often the first thing a phone user sees, before the website. How to work on that ranking is explained in our guide to local SEO.
The third is intent. For "assicurazione auto" (car insurance) in Italy the mobile SERP was made of quote pages and the desktop one of articles. In the UK the gap was smaller but real: on desktop the top results were the main quote pages of insurers and comparison sites, while on mobile they were a mix of insurers' pages, lists of providers, a guides hub and a premium calculator. Before deciding which page to write for a search, look at both versions: the search intent Google assigns to a query may not be the same on the two devices.
How to measure mobile SEO in Search Console
The Performance report in Google Search Console lets you filter clicks, impressions and average position by device. Google's own help page points out that search results vary with the time, place, device and history of the person searching, which is why the position you see yourself doesn't match the one in the report.
- Open Performance, choose the Devices tab and compare clicks and average position for Mobile and Desktop over the last three months.
- Go back to Queries, add the filter Device: Mobile, then repeat with Desktop. Queries where the average position differs by more than three places are the ones to check by hand, on a phone and on a computer.
- In the Core Web Vitals report, start with the Mobile group: according to the Web Almanac data, that's where you'll find the most URLs outside the thresholds.
- If you use a rank tracking tool, check which device it measures on and, if you can, track both.
This device comparison is one of the first checks we run when a client comes to us saying "nobody can find us on their phone". If you'd like us to do it on your site, it's part of our SEO consultancy work, along with the rest of technical SEO.
Mobile SEO statistics you'd better not quote
While preparing this page we checked where the most repeated figures in mobile SEO guides come from. Three examples. SEO.com's guide to mobile SEO says "58% of Google searches come from smartphones" and "40% of adults use voice search daily", citing other pages on the same site rather than a study. The first figure, however you measure it, describes the American market. On Google UK, the AI Overview for "mobile seo" on 10 October 2026 opened with "over 60% of online searches happen on mobile devices", a global-sounding figure that, as we've seen, doesn't match the UK page-view split. And the Nightwatch guide, on page one of Google UK that same day, still lists FID among the Core Web Vitals and Google's Mobile-Friendly Test among the tools, two and a half years after FID was replaced and nearly three after the test was retired.
For a UK site, the useful numbers are the ones you can reproduce: the mobile and desktop split in Search Console, the Core Web Vitals report by device, and the SERP as seen from a phone for your ten main searches.
Mobile SEO FAQs
Mobile-first indexing is how Google indexes pages: it reads the smartphone version. Mobile SEO is broader and also covers speed on a phone, usability, how the page looks in the mobile SERP and rankings for searches made on a smartphone, which can return different results from the same searches on desktop.
Yes, often. In our test of 20 searches in Italy on 3 October 2026, the top 10 organic results had on average 2.6 URLs in common between smartphone and desktop, and none at all in 8 searches. A smaller check on Google UK found zero URLs in common for 'car insurance' and 'plumber manchester'. In Search Console you can see your site's average position split by device.
No, because it no longer exists. Google retired the Mobile-Friendly Test and the Mobile Usability report on 1 December 2023. To check a page, use URL Inspection in Search Console, Lighthouse and PageSpeed Insights.
It depends how you read the question. Google measures them separately for mobile and desktop, and sites pass them less often on mobile: 48% against 56% in 2025, according to the Web Almanac. Google says Core Web Vitals are used by its ranking systems, but they mostly count when pages are equal on content and authority.
No. Responsive design, with the same URL and the same HTML for every device, is the setup Google recommends, and none of the 49 Italian pages in our mobile-first indexing test still used a separate mobile site. AMP is no longer required even for the Top Stories carousel.
The most reliable way is to search from a real phone, in a private window. Otherwise you can use device mode in Chrome DevTools or a rank tracking tool set to mobile, bearing in mind that personalisation and location will still change the result.
The report hardly anyone reads
Many SEO reports show a single position per keyword, without saying which device it was measured on. The person receiving the report checks it by searching on their own phone. When the two numbers don't match, the usual conclusion is that the report is wrong. In our sample the difference is the norm: in 16 of 20 searches the first result on a phone wasn't the same URL as the first on a computer, and in the UK check it differed for both "car insurance" and "plumber manchester". A report that shows position, clicks and local pack separately by device heads off a lot of arguments about results, and often reveals an opportunity the average was hiding.
