This guide explains what SpeedSize optimizes, what KPIs matter, and how to correctly measure results after integration. Many customers naturally check PageSpeed Insights (PSI), Shopify reports, Lighthouse, or simply rely on subjective impressions of site speed. While these checks are common, it’s important to understand how our service works and what to expect.

This article contains the following topics:

1. What SpeedSize Does

Our innovative technology allows you to load media up to 10x faster while preserving 100% visual quality. By employing AI Neuroscience technology, SpeedSize intelligently analyzes each media file, removes unperceivable data, and delivers optimized variations tailored to each device and browser.

AI-Driven Media Optimization

Compresses and enhances images and videos without visible quality loss.
When integrated into a CMS like Shopify, SpeedSize improves image quality compared to Shopify’s native compression while significantly reducing file sizes.
Automatically delivers media in modern, efficient formats (AVIF, WebP for images, WebM for videos).
Allows you to upload the highest-quality originals without worrying about performance.

Ultra-Fast CDN

Uses a globally distributed CDN with edge computing. Media is delivered from the nearest server, reducing latency and TTFB.
Improved cache/miss ratio ensures faster repeat visits and smoother browsing.

Adaptive Media Delivery

Serves the best version of an asset for each device and browser.
Automatically adjusts resolution and size (e.g., smaller versions for mobile), reducing bandwidth and load times.
Uses Client Hints and srcset/sizes for precise responsive delivery.

Core Web Vitals (CWV) – LCP Focus

While CWV are not SpeedSize’s KPIs, we can significantly improve Largest Contentful Paint (LCP) if the LCP element is an image or video:

We directly reduce Resource Load Duration for LCP media.
Example: CrUX data demonstrating that load duration was cut by more than three times with SpeedSize enabled.

Important: Overall LCP also depends on factors we do not control, such as:

Resource Load Delay
Element Render Delay
Site TTFB

Other CWVs like INP, CLS, and FCP are not influenced by SpeedSize since they depend on site code and layout.

2. What SpeedSize Does Not Do

We don’t optimize your site’s code (JavaScript, CSS, or third-party apps).

SpeedSize focuses entirely on media files, removing the media bottleneck that slows down most modern websites. Images and videos are often the heaviest assets on a page, so optimizing them is one of the most impactful steps you can take. By solving this bottleneck, SpeedSize ensures your media loads faster, sharper, and more efficiently.

That said, overall site performance also depends on other factors — JavaScript, CSS, and third-party apps can still cause delays. These remain outside our scope and may need to be addressed separately for full performance optimization.

We don’t directly improve all Core Web Vitals.

The only CWV we can partially influence is LCP — and only if the LCP element is an image or video.
In those cases, we reduce Resource Load Duration by delivering a smaller, faster asset.
But LCP can also be affected by render-blocking JavaScript or CSS, lazy loading, browser rendering delays, and overall site TTFB — all outside our control.

Metrics like INP, CLS, and FCP depend entirely on site code, layout, and scripts.

We don’t guarantee higher PageSpeed Insights (PSI) or Lighthouse scores.

PSI and Lighthouse provide synthetic lab data, simulating loading under fixed conditions. While useful for debugging, they don’t always reflect real-world performance.
These tools also flag issues unrelated to media, such as heavy JavaScript or render-blocking CSS.

This is why we recommend focusing on field data (real user data, e.g., CrUX) when evaluating LCP and other CWVs. Since CrUX aggregates 28 days of data, the most accurate results appear after about a month of SpeedSize being live.

3. How to Correctly Measure SpeedSize Impact

Media Quality and Speed

The most immediate way to see SpeedSize's impact is by looking at your media directly. After integration, images and videos will:

Load faster (smaller file sizes, quicker delivery).
Look better (optimized without over-compression).

You can compare “before vs. after” directly in your browser or with DevTools, checking file size reductions and improved timings.

We also provide before/after comparisons of key assets right after integration. These are the clearest proof of SpeedSize’s value.

Field Data (Real User Monitoring)

The most reliable way to measure SpeedSize impact is using real user monitoring (RUM) data. This reflects actual user experience across devices, locations, and network conditions.

You can measure SpeedSize impact using:

Chrome User Experience Report (CrUX)
New Relic Browser Monitoring

Chrome User Experience Report (CrUX)

CrUX collects data from actual Chrome users over a rolling 28-day period.

Improvements in LCP and media delivery are typically visible after ~30 days of SpeedSize being live.
CrUX reflects real user experience, not just lab simulations.
Useful for validating long-term production improvements.

New Relic Browser Monitoring

New Relic allows measuring SpeedSize impact using real-time user traffic. Unlike CrUX, it provides immediate visibility into performance changes.

With New Relic, you can measure:

LCP Resource Load Duration (primary SpeedSize KPI)
Total LCP impact
Before vs. after comparisons
A/B testing (SpeedSize vs existing CDN)

Particularly useful for Enterprise PoCs, A/B testing scenarios, faster performance verification (hours/days instead of weeks). For detailed setup instructions, see our Measuring SpeedSize with New Relic guide.

Lab Data (PageSpeed Insights)

PSI is helpful for debugging but should not be the sole measure of SpeedSize's impact.
Use it to confirm that media-related issues (e.g., "Improve images delivery") are resolved.
When analyzing LCP in PSI, focus on Load Time of the LCP image or video — the area SpeedSize directly influences.
Remember: PSI often flags other issues (JavaScript, render-blocking CSS, third-party scripts) that are outside SpeedSize's scope.

4. Recommendations for Maximizing SpeedSize Impact

Prioritize LCP Media

Avoid lazy loading for hero images and above-the-fold content.
Use loading="eager" and fetchpriority="high" on your LCP image to ensure it loads early.

Proper Sizing and Responsive Delivery

Keep intrinsic-to-rendered width ratio close to ~1.5–2.0x for best performance.
Confirm Client Hints are enabled and your server sends the correct response headers.
Use srcset and sizes so the browser selects the optimal image version.

Preload Critical Media

Preload videos (via the preload attribute) or key images when necessary to reduce delays and ensure smoother playback or rendering.

More tips can be found in our Advanced Integration & Best Practices guide.

Key Takeaways

SpeedSize solves the media bottleneck: your images and videos are now optimized, sharp, and delivered as fast as possible.
We cannot control site code, third-party apps, or non-media performance issues.
The most meaningful KPIs for SpeedSize are media quality, reduced file sizes, and faster delivery.
When media is the LCP, we also significantly improve its Resource Load Duration.
For evaluation, always rely on real-user data (RUM), not just synthetic PSI tests.