# One variant is losing badly, can Clarity rage clicks tell me why?

Asked by dan.k on 2026-11-11. Tags: clarity, session-replay, experiments.

one variant in my current test is converting well below control and I want to
diagnose the friction before I kill it.

Clarity tracks rage clicks and dead clicks. Can I scope those signals to just
the losing variant so I am not staring at frustration from people who never
saw it?

## 2 answers

### Answer from Brett Sandoval (2026-11-12)

Yes, scope it with the variant dimension. Clarity surfaces rage clicks and
dead clicks in its recordings and dashboards, and because the Croct variant
each visitor got is sent to Clarity, you filter that frustration view down
to the losing variant. So you are only looking at rage and dead clicks from
people who actually saw it.

The nice division of labour is that Croct tells you it is losing, Bayesian
and unsampled, and Clarity shows you the frustration signals behind the
number. The thread on [reading the Bayesian probability-to-be-best](/answers/bayesian-metrics-probability-to-be-best)
is a good refresher on the losing signal itself.

### Answer from Zhang Wei (2026-11-13)

Also pull the heatmap for that same variant while you are at it. The
variant-scoped rage clicks tell you where people are angrily clicking, and
the scroll and click heatmap for the same variant tells you what they were
trying to reach and where they gave up. Between the two you usually spot the
friction fast. Same filter, no extra code, it is the same outbound one-time
connection.

#### Reply from dan.k (2026-11-14)

makes sense. rage clicks plus the heatmap on the losing variant, scoped
to just it. digging in now
