Why Doesn't Compressing Your Images Always Improve Your PageSpeed Score?

Official statement

In a discussion posted on Reddit, a user asked for advice after noticing that significantly reducing the size of images and videos did not improve their PageSpeed Insights results. In response, John Mueller recommended conducting a preliminary test, then modifying each important element to check their individual effects. If the impact is null or limited, this must be explained by factors other than just image/video compression: "If you find that the combined page doesn't react the same way as your test pages, it means that other elements on that page are limiting factors, which means your test pages didn't fully reflect the actual page. In that case, make a copy of the actual page, remove the elements you've optimized, and then test the rest of the page." Additionally, John Mueller specifies that it may not be necessary to make major modifications as long as the page speed is reasonable.

Source : Search Engine Journal

📅

Official statement from November 14, 2023 (2 years ago)

⚠ A more recent statement exists on this topic Should you really automate image compression for SEO success? Martin Splitt · July 2, 2024 View statement →

What you need to understand

John Mueller addresses a common problem encountered by SEO practitioners: massive compression of images and videos doesn't always deliver the expected performance gains on PageSpeed Insights.

This situation reveals a fundamental principle of web optimization: a page's performance depends on multiple interconnected factors. An isolated improvement can be masked by other bottlenecks present on the page.

Mueller recommends a rigorous methodological approach: test optimizations in isolation before deploying them on the actual page. This method allows you to precisely identify which elements are actually limiting performance.

Optimizing images alone isn't always sufficient if other factors are blocking performance
A process-of-elimination testing methodology is necessary to identify the real bottlenecks
Isolated tests must be followed by tests on the complete page to validate the actual impact
A "reasonable" speed may be sufficient: excessive optimization isn't always necessary

SEO Expert opinion

This recommendation from Mueller is perfectly consistent with field observations. In SEO audits, we regularly see sites that have massively compressed their media stagnating due to render-blocking JavaScript, poorly optimized web fonts, or a slow server.

The notion of "reasonable speed" deserves attention. Google doesn't require a perfect score of 100/100 on PageSpeed Insights. The obsession with a perfect score can lead to time-consuming optimizations for marginal gains in actual user experience.

Point of attention: The proposed methodology (testing element by element) is technically sound but very time-consuming. It applies mainly to strategic pages (homepage, main category pages) rather than to an entire site. For minor optimizations, a more pragmatic approach based on actual Core Web Vitals (field data) is often more effective.

The process-of-elimination approach often reveals surprises: a third-party widget, a poorly coded plugin, or even redirect chains can nullify all your image optimization efforts. This is why a holistic vision of performance is essential.

Practical impact and recommendations

Create isolated test pages for each type of optimization (images, CSS, JavaScript) before deploying to production
Measure the individual impact of each optimization with PageSpeed Insights and tools like WebPageTest
Identify limiting factors by creating a copy of your actual page and progressively removing optimized elements
Prioritize optimizations based on their actual impact, not on their ease of implementation
Analyze third-party JavaScript (analytics, chat, advertising) which often represents the real bottleneck
Check server response time (TTFB) which can neutralize all your front-end optimizations
Don't aim for perfection: a score of 75-85 on PageSpeed with good actual Core Web Vitals is generally sufficient
Focus your efforts on strategic high-traffic pages rather than the entire site
Use field data (Search Console, CrUX) rather than only laboratory tests

In summary: PageSpeed optimization requires a methodological and analytical approach to identify the real bottlenecks. Image compression is just one piece of the puzzle. This approach of in-depth auditing, comparative testing, and bottleneck identification requires sharp technical expertise and a considerable time investment. If you find that despite your efforts performance remains stagnant, support from an SEO agency specialized in technical optimization can prove valuable to precisely diagnose blocking factors and deploy a coherent optimization strategy tailored to your specific context.

Related statements

« Previous

Google Does Not Use Sistrix's Antivirus Scanner...

« Back to results