Behavioral Scores
Every session in recapt gets behavioral scores from 0 to 1. These scores tell you at a glance whether users are having a good experience or struggling.
How Scores Work
Behavioral scores are calculated automatically for every session. A score of 0 means no signal detected, while 1 means the strongest possible signal. Most sessions fall somewhere in between.
Scores are updated in real-time as the session progresses, so you can see how user experience changes throughout their visit.
Importantly, scores incorporate micro-signals: subtle behaviors like tiny scroll adjustments, brief hesitations, and small cursor movements that happen too fast for the human eye to catch when watching recordings. These micro-behaviors are often invisible during manual review but contribute meaningfully to the overall score.
Frustration Score
The frustration score measures how annoyed or blocked a user appears to be.
What triggers high frustration:
- Rapid, repeated clicks on the same element (rage clicks)
- Clicking on elements that don't respond (dead clicks)
- Quick scroll reversals (scrolling down then immediately back up)
- Repeated backtracking to previous pages
- Multiple failed form submissions
- Micro-frustrations: small, rapid interactions that individually seem normal but accumulate into a frustration pattern
Example scenarios:
| Score | What it looks like |
|---|---|
| 0.0 - 0.2 | Normal interaction, no signs of frustration |
| 0.3 - 0.5 | Some friction: a few repeated clicks or minor hesitation |
| 0.6 - 0.8 | Clear frustration: rage clicks, scroll thrashing, or repeated failures |
| 0.9 - 1.0 | Severe frustration: user is clearly blocked or struggling significantly |
Confusion Score
The confusion score measures how lost or uncertain a user appears to be.
What triggers high confusion:
- Long pauses before taking action (hesitation)
- Erratic, wandering mouse movements
- Taking indirect paths to reach a destination
- Visiting the same pages multiple times
- Starting and abandoning forms or flows
- Micro-hesitations: brief pauses and small cursor adjustments that suggest momentary uncertainty
Example scenarios:
| Score | What it looks like |
|---|---|
| 0.0 - 0.2 | User knows where they're going, direct navigation |
| 0.3 - 0.5 | Some uncertainty: exploring options or minor hesitation |
| 0.6 - 0.8 | Clearly confused: wandering, backtracking, long pauses |
| 0.9 - 1.0 | Very lost: can't find what they need, circular navigation |
Confidence Score
The confidence score is the inverse of frustration and confusion. It measures how smoothly and purposefully a user is navigating.
What indicates high confidence:
- Direct navigation to goals
- Steady, purposeful mouse movements
- Quick, decisive clicks
- Completing tasks without backtracking
- Smooth form completion
A high confidence score (0.7+) typically indicates a user who knows what they want and is finding it easily. This is what you want to see.
Using Scores Effectively
Prioritizing Sessions to Review
Instead of watching random sessions, filter by behavioral scores to find the ones that matter:
- High frustration (0.6+): Users who hit problems you need to fix
- High confusion (0.6+): Users who couldn't find what they needed
- Low confidence on key pages: Potential UX issues on important flows
Tracking Trends Over Time
Monitor average scores across your site to spot emerging issues:
- Rising frustration on a specific page? Something changed.
- Confusion increasing after a redesign? Navigation might need work.
- Confidence dropping on checkout? Investigate immediately.
Setting Up Alerts
Configure alerts based on score thresholds to catch issues early:
- Alert when frustration exceeds 0.7 on critical pages
- Alert when confusion spikes above baseline
- Alert when confidence drops significantly
Scores vs. Events
Behavioral scores and events (like rage clicks) are related but different:
- Events are discrete occurrences: "a rage click happened here"
- Scores are continuous measures: "this session had 0.65 frustration overall"
Use events to pinpoint specific moments. Use scores to understand the overall experience and compare sessions.