Website Speed Testing Tools Compared: PageSpeed Insights, GTmetrix, and WebPageTest

Run your website through three different speed testing tools and you’ll likely get three different scores. PageSpeed Insights might give you 62, GTmetrix an A grade, and WebPageTest a different breakdown again. This inconsistency confuses business owners and developers alike — but the differences aren’t arbitrary. Each tool measures speed from a different angle, uses different testing infrastructure, and emphasises different aspects of performance. Understanding what each tool actually measures helps you focus your optimisation efforts on what matters most.

At Xpose, we use all three tools as part of our performance audits for clients across Norwich and the wider UK. In this guide we explain what PageSpeed Insights, GTmetrix, and WebPageTest each measure, how to read their output, where they agree and where they diverge, and which tool gives the most reliable signal for each type of decision you need to make about your website’s speed.

PageSpeed Insights: Google’s Official Performance View

PageSpeed Insights (PSI) is the tool most businesses encounter first because it’s made by Google and its scores are perceived — rightly or wrongly — as directly influencing search rankings. PSI runs your URL through Lighthouse, Google’s open-source auditing engine, and produces a score from 0 to 100 across four categories: Performance, Accessibility, Best Practices, and SEO. The Performance score is the one most people focus on. Critically, PSI also pulls in real-world data from Chrome users who have visited your site, displayed in the Core Web Vitals section. This field data — measuring Largest Contentful Paint, Cumulative Layout Shift, and Interaction to Next Paint — is what Google actually uses as a ranking signal, making it the most important section on the page.

The main limitation of PSI’s lab score is that it simulates a throttled mobile device on a slow 4G connection, which often produces low scores even for genuinely fast sites. A score of 50 in PSI doesn’t mean your site feels slow to your actual visitors — it means it’s slow on a simulated mid-range Android phone on a throttled network. Always read the field data alongside the lab score. If your Core Web Vitals field data shows green across all three metrics, your real-world users are having a good experience regardless of the lab score. Focus your optimisation efforts on metrics where the field data shows amber or red.

GTmetrix: Detailed Waterfalls and Historical Tracking

GTmetrix runs tests from real servers in multiple locations and produces a detailed waterfall chart showing every single resource your page loads, in the order it loads, with timing for each request. This waterfall view is GTmetrix’s biggest advantage over PSI — it lets you see exactly which resource is slowing your page down, whether that’s a slow third-party script, an unoptimised image, or a blocking stylesheet. GTmetrix also lets you test from multiple geographic locations, which is useful for diagnosing whether slow load times are a server location issue affecting users in particular regions.

GTmetrix grades pages with a letter grade (A to F) based on a composite of performance metrics, and it lets you track your score over time if you create an account. This historical view is particularly useful for monitoring whether a recent change improved or degraded performance. One important caveat: GTmetrix’s default test runs on a desktop connection, which will produce faster results than PSI’s mobile-throttled lab test. Always switch to a mobile test profile and a relevant geographic location when comparing results with PSI. At Xpose, we use GTmetrix primarily for its waterfall analysis when diagnosing specific bottlenecks.

WebPageTest: The Most Detailed and Reliable Tool

WebPageTest (webpagetest.org) is the most powerful of the three tools and the one professional performance engineers rely on most heavily. It runs tests from real browsers on real hardware in data centres around the world, gives you full control over connection speed and browser type, and provides metrics not available in the other tools — including Start Render, Speed Index, Time to First Byte (TTFB), and detailed filmstrip views showing exactly how the page renders frame by frame. The filmstrip is particularly useful for identifying layout shifts and render-blocking resources that affect perceived load speed even when overall metrics look reasonable.

WebPageTest also supports scripting, allowing you to test authenticated pages, multi-step interactions, and repeat-view performance (which shows whether your caching strategy is working). Its "Opportunities and Experiments" section suggests specific improvements and even runs automated tests to quantify exactly how much each optimisation would improve your score. The tradeoff is complexity — WebPageTest produces more data than most business owners know what to do with. Our recommendation at Xpose: use PSI first for a quick health check and to monitor Core Web Vitals field data; use GTmetrix when you need to diagnose a specific resource bottleneck; use WebPageTest for deep performance audits and before-and-after comparisons after major changes.