Accuracy

Why our score won't match PageSpeed Insights

If Speed Doctor gives a different score than Google PageSpeed Insights, that's expected — not a bug. Both run Lighthouse, but where, how and how often they run it differs. Here's exactly why the numbers diverge.

Short version
Performance scores are a lab measurement. Change the machine, the network, the location, the Lighthouse version, or the number of runs, and the score moves. Use Speed Doctor to find what to fix and to track your own trend — use it alongside PSI, not as a clone of it.

1. Different machine and CPU speed

Lighthouse's score depends heavily on CPU performance. Google runs PSI on its own standardised infrastructure; Speed Doctor runs on whatever server (or laptop) hosts your worker. A faster or slower CPU shifts metrics like Total Blocking Time and, with it, the overall score.

2. Different network location

The distance between the testing machine and your website's servers changes latency and download times. PSI tests from Google's data centres; your worker tests from its own region. That alone can move Time to First Byte and Largest Contentful Paint by hundreds of milliseconds.

3. Different throttling model

Speed Doctor uses Lighthouse's simulated throttling for fast, consistent results. PSI uses its own calibrated throttling tuned to a reference device. Different throttling assumptions produce different timings even for the identical page.

4. Different Lighthouse & Chrome versions

Lighthouse changes its scoring weights and metric definitions between versions. If Speed Doctor and PSI are on different Lighthouse or Chrome builds, the same page can score differently purely because the rules changed.

5. Single run vs. medians & variance

Lighthouse results naturally vary run to run — ads, A/B tests, lazy content and third-party scripts all introduce noise. Speed Doctor reports a single run for speed; treat one number as a snapshot, and look at the trend across several runs rather than a single absolute figure.

6. Lab data vs. real-user (CrUX) field data

PSI also shows field data from the Chrome User Experience Report — metrics from real visitors over the last 28 days. Speed Doctor produces lab data from a controlled test. These answer different questions: field data is what users felt; lab data is what you can reproduce and debug.

7. Speed Doctor adds its own analysis

Beyond Lighthouse, Speed Doctor runs six DOM detectors (oversized images, render-blocking scripts, heavy third parties, DOM size and more) and AI root-cause correlation. These power the issue list and fixes — extra signal that PSI doesn't surface, but which doesn't change the Lighthouse score itself.

How to use both tools well

  • Use PageSpeed Insights for the official Google score and real-user field data.
  • Use Speed Doctor to diagnose root causes, get developer-ready fixes, and track changes over time.
  • Compare like with like: same device mode, same page state, several runs.
  • Focus on the issues and metrics you can act on, not a single headline number.