PageSpeedIQ
Back to PageSpeedIQ
Engineering Transparency•Updated October 2026

Audit Methodology & Standards

At PageSpeedIQ, we believe that web performance audits must be transparent, repeatable, and grounded in verifiable data. Here is the exact specification of how our platform executes audits, controls network environments, and interprets results.

Deterministic Lab

Clean-room cloud instances eliminate browser extensions, ad-blockers, and local computer CPU bias.

Honest Field Data

We clearly separate synthetic lab results from Google CrUX 28-day real-user telemetries.

Zero Simulated Fakes

Live audits return genuine Google PageSpeed data. If upstream fails, we report the exact error, never simulated fake scores.

1. Audit Engine & Infrastructure

Google Lighthouse v13.5 Execution

2. Environmental Emulation Specifications

Throttling profiles for mobile and desktop runs

Mobile Profile (Moto G Power)
  • • Viewport: 412 × 823 px (DPR 1.75)
  • • Network: Simulated Slow 4G
  • • Download Throughput: 1,638 Kbps
  • • Upload Throughput: 750 Kbps
  • • Round-Trip Latency: 150 ms
  • • CPU Throttling: 4x slowdown
Desktop Profile
  • • Viewport: 1,366 × 768 px (DPR 1.0)
  • • Network: Fast Wired/Wi-Fi
  • • Download Throughput: 10,000 Kbps
  • • Upload Throughput: 10,000 Kbps
  • • Round-Trip Latency: 40 ms
  • • CPU Throttling: 1x (no slowdown)

3. Field Data (CrUX) Integration

How we report Chrome User Experience metrics

Missing Field Data Transparency:

If a website does not meet Google minimum traffic volume, field data is marked as “Field Data Not Available (Insufficient CrUX Traffic)”. We never invent or extrapolate field numbers.

4. Repeatability & Cache Windows

Preventing API quota exhaustion & ensuring test integrity

Independent Platform Disclaimer