Engagement Analytics
Understand how users interact with your website and correlate engagement with performance.
VitalSentinel tracks user engagement metrics and correlates them with performance data, helping you understand how site speed affects user behavior.
This page is part of RUM Monitoring and lives at Domain → RUM → Engagement.
How It Works
When engagement tracking is enabled (default), the RUM script captures:
- Scroll behavior - How far users scroll down the page
- Click interactions - Rage clicks and clicks on interactive elements
- Time on page - Active engagement time
- Form interactions - Form abandonment and completion
What You See
The Engagement page is built around a single performance × engagement correlation chart. You pick one performance metric on one axis and one engagement metric on the other, and the chart shows how they relate across your real users.
Summary Cards
Above the chart you'll find:
- Total Sessions in the selected window
- Good Performance – the share of sessions where the chosen performance metric was in the "good" range
- Average value of the chosen engagement metric
Correlation Chart
Bar chart of session distribution across performance buckets, with a line overlay of the selected engagement metric per bucket.
Performance metric (selectable): LCP, INP, CLS, FCP, TTFB.
Engagement metric (selectable):
| Metric | Description |
|---|---|
| Scroll Depth (25%, 50%, 75%, 90%, 100%) | Share of sessions that scrolled at least this far |
| Clicks per Session | Average clicks per session |
| Bounce Rate | Single-page sessions |
| Rage Click Rate | Sessions with rapid frustrated clicking |
| Form Abandonment Rate | Sessions that started but did not submit a form |
| Time on Page | Average active engagement time |
| Attention Score | Combined engagement quality score |
| Time to First Interaction | Average time until the first interaction |
Typical patterns:
- Slower LCP → higher bounce rate
- Better INP → more clicks per session and lower rage-click rate
- Lower TTFB → longer time on page
Rage Click Detection
Rage clicks are detected when users click rapidly in frustration:
- 3 or more clicks on the same element
- Within a 2-second window
High rage click rates indicate:
- Broken links or buttons
- Slow-responding UI elements
- Confusing interface design
Finding Rage Click Sources
- Filter by high rage click rate
- Check which pages have issues
- Review click targets
- Fix unresponsive elements
Form Analytics
The RUM script tracks form focus, field changes, submissions, and abandonment. In the dashboard this surfaces as the Form Abandonment Rate engagement metric – the share of sessions that started filling a form but left without submitting. Select it in the correlation chart to see how performance relates to form abandonment.
Reducing Form Abandonment
- Simplify form fields
- Add progress indicators
- Improve form validation
- Optimize form load time
Filtering Data
The Engagement page shares the standard RUM filter bar:
- Date range – any window, with an optional compare-to-previous-period toggle
- Device – Mobile or Desktop
- Advanced filter – narrow by URL path, country, or browser
See RUM Monitoring for details on the filter bar.
Use Cases
Measuring Content Engagement
Track scroll depth to understand:
- Is content engaging enough?
- Do users read full articles?
- Where do users drop off?
Improving Conversion
Correlate performance with actions:
- Faster pages = more sign-ups?
- Page speed affecting purchases?
- Performance impact on engagement?
Identifying UX Issues
Find problems via:
- High rage click rates
- Low scroll depth despite long content
- High form abandonment
- Low time on page
Best Practices
Set Engagement Goals
Define targets for:
- Minimum scroll depth for key pages
- Maximum acceptable bounce rate
- Target time on page
Monitor Trends
Watch for changes:
- Engagement drops after deployments
- Seasonal patterns
- Device-specific issues
Correlate with Performance
Use correlation data to:
- Justify performance improvements
- Quantify speed impact on business
- Prioritize optimization work