When using LoadRunner VuGen TrueClient for browser-based performance testing, understanding event-based timing is critical. In a VIN Search transaction, a mismatch was observed between tool metrics and actual user experience after backend optimization.
Problem Observed
Before Optimization:
- Actual VIN Search time: ~9 seconds
- VuGen TrueClient reported: ~9 seconds
- No discrepancy
After Optimization:
- Actual VIN Search time: 2–3 seconds
- VuGen TrueClient reported: ~9 seconds
- Apparent regression in reports despite real improvement
Why This Happened
TrueClient does not inherently know when a business transaction is complete. It relies on configured start and end events.
By default, the protocol uses “Network Completed” as the end trigger, which includes:
- VIN search request
- Server processing
- HTML rendering
- JavaScript execution
- CSS, images, fonts, and other resource loading
- Post-load background network calls
For highly dynamic web applications, this can significantly inflate reported response times.
What Was Changed
Start Event:
- Triggered just before pressing ENTER after VIN entry
End Event:
- Set to “Action Completed” when VIN details were visible on screen
Excluded:
- All non-critical post-render background activity from timing
Outcome
- VuGen reported response times aligned with manual stopwatch measurements (2–3 seconds)
- Stakeholders regained confidence in performance metrics
- Test results accurately reflected real user experience
Technical Takeaway
Always customize TrueClient start and end events to match business KPIs instead of relying on default network timing triggers.
