You run your site through PageSpeed Insights. Desktop score: 92. Green across the board. You feel good. Then you check the mobile tab. Score: 41. Everything is red.
This is the most common performance disconnect I see. And it is the one that matters most, because Google uses mobile-first indexing. It crawls the mobile version of your site. It evaluates mobile Core Web Vitals. The desktop score, as reassuring as it is, has no direct impact on how Google assesses your performance for ranking.
Why mobile scores are always worse
The performance gap between desktop and mobile is not a bug. It reflects real differences in how the two environments process your page.
CPU power. The Lighthouse mobile test simulates a Moto G Power, which has a CPU roughly 4x slower than a modern laptop. Every JavaScript operation that takes 50ms on your MacBook takes 200ms on the simulated mobile device. This is why framework-heavy sites with lots of JavaScript see dramatic score differences between desktop and mobile.
Network speed. The mobile test throttles the connection to simulate 4G speeds, which are significantly slower than the broadband connection used in the desktop test. Large images and video that download in 1 second on desktop take 3-5 seconds on the simulated mobile connection.
Screen size. Mobile screens are smaller, which changes which elements are in the viewport and affects LCP measurement. An image that is below the fold on desktop might be the LCP element on mobile. Or vice versa.
Touch interactions. Mobile INP captures touch-based interactions, which can behave differently from mouse clicks. Scroll-jacking, swipe handlers, and touch-delay JavaScript all affect mobile INP specifically.
Why mobile is what Google cares about
Since 2019, Google has used mobile-first indexing for all websites. This means:
- Googlebot crawls the mobile version of your site by default.
- Core Web Vitals assessments are based on mobile field data from the Chrome User Experience Report (CrUX).
- Content on mobile is what gets indexed. If content is hidden on mobile but visible on desktop, Google may not index it.
Your desktop score is irrelevant to Google's ranking algorithms. A site with a 95 desktop score and a 35 mobile score is evaluated at 35 for ranking purposes.
Want a website that turns visitors into customers, not just compliments?
Book a 15-min introWhat field data actually shows
For most websites, the CrUX data skews heavily toward mobile. Depending on your industry:
- Local service businesses: 65-80% mobile traffic
- E-commerce: 55-70% mobile traffic
- SaaS B2B: 40-60% mobile traffic (skews more desktop than other industries)
- Content/media: 60-75% mobile traffic
This means your field data, which is what Google uses for ranking, is dominated by mobile user experiences. Even if your desktop field data is excellent, poor mobile field data drags down your overall Core Web Vitals assessment.
You can check the device split in Google Search Console under Performance, then filter by device. Or look at your analytics data for the mobile vs desktop traffic split.
Mobile-specific optimizations
Some optimizations matter more on mobile than desktop:
JavaScript is the top priority. The 4x CPU difference means JavaScript performance improvements have an outsized impact on mobile scores. Code-splitting, deferred loading, and trimming unused JS all disproportionately help mobile. I covered this in my post on making heavy framework sites fast.
Image sizing matters more. On mobile, serving a 1440px-wide image that displays at 375px wide wastes more relative bandwidth than the same waste on desktop. Use responsive images (srcset) to serve appropriately sized files for each viewport width.
Touch target sizing. While not a speed metric, small touch targets force users to zoom and re-tap, which increases interaction time and affects perceived performance. Ensure buttons and links are at least 48x48 pixels on mobile.
Above-the-fold content is different. The smaller viewport means less content is above the fold. What counts as LCP on mobile may be different from desktop. Test both and optimize the LCP element for each breakpoint.
How to test mobile performance properly
Do not rely only on Chrome DevTools emulation. Device emulation in Chrome throttles the CPU and network but does not replicate real mobile hardware behavior. It is a useful approximation, not a definitive test.
Use WebPageTest with a mobile profile. WebPageTest lets you test from real device profiles and real geographic locations. Set the test to a Moto G4 on 4G to get realistic mobile performance data.
Test on a real device. If possible, test on a mid-range Android phone (Samsung Galaxy A series, Pixel 4a, or similar). These represent what your actual mobile visitors use. The performance gap between your MacBook and these devices is real and significant.
Check field data in Search Console. The Core Web Vitals report in Search Console separates mobile and desktop field data. This is the authoritative source for understanding how Google evaluates your mobile performance.
The prioritization framework
When time is limited, prioritize mobile fixes in this order:
- LCP on mobile. Ensure the hero image is optimized, eagerly loaded, and appropriately sized for mobile viewports.
- INP on mobile. Audit JavaScript weight and defer non-critical scripts. The CPU difference makes this critical on mobile.
- CLS on mobile. Check for images without dimensions and elements that shift on smaller screens.
- Then desktop. Once mobile scores are green, optimize desktop. The work often overlaps because many fixes improve both.
This order ensures you address the ranking-relevant metrics first. A site with green mobile Core Web Vitals and yellow desktop metrics is in a better position than the reverse.
If your mobile scores are dragging down your search performance while your desktop scores look fine, that gap is exactly what a mobile-focused speed audit addresses. The fix is usually JavaScript and image optimization targeted specifically at mobile behavior.