Key finding: A/B testing tool scripts range from 4KB (ExperimentHQ) to 203KB (Optimizely). On 3G networks, this translates to 45ms vs 1,270ms load time. Heavy scripts directly hurt the conversions you're trying to improve. Choose tools under 10KB for optimal performance.
Script Size Benchmark (December 2025)
| Rank | Tool | Size (KB) | Gzipped | 3G Load | 4G Load |
|---|---|---|---|---|---|
| 1 | ExperimentHQ | 4.2KB | 1.8KB | 45ms | 15ms |
| 2 | PostHog | 28KB | 9KB | 180ms | 60ms |
| 3 | GrowthBook | 12KB | 4KB | 90ms | 30ms |
| 4 | Convert | 82KB | 28KB | 520ms | 175ms |
| 5 | AB Tasty | 118KB | 38KB | 750ms | 250ms |
| 6 | VWO | 156KB | 48KB | 980ms | 330ms |
| 7 | Optimizely | 203KB | 62KB | 1270ms | 425ms |
Last updated: December 2025. Methodology below.
Performance Impact
Script size directly affects conversions:
- • Every 100ms delay = ~1% conversion loss (Amazon research)
- • 200KB script on 3G = 1,200ms delay = ~12% conversion loss
- • The irony: Tools meant to improve conversions actually hurt them
Methodology
Measurement
Chrome DevTools Network tab, cache disabled, repeated 10 times
Network conditions
Tested on 3G (1.6 Mbps), 4G (9 Mbps), and broadband (50 Mbps)
Script scope
Main snippet only (not including dependencies or tracking pixels)
Update frequency
Re-measured monthly to track changes
Download Raw Data
Full benchmark data available as CSV for citation in your research or blog posts.
Choose Lightweight Tools
ExperimentHQ is 40-50x smaller than enterprise tools. Same A/B testing capabilities, minimal performance impact.