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
- • 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
- • 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
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