Error Analysis uses approximation to optimize performance and provide faster results.
In Error Analysis, the number of Sessions with error or Sessions with Rage Clicks in the Errors table is approximated. The KPI calculations displayed on the error table may therefore have a % margin of error, based on the size of the data set as indicated below.
How Error Analysis approximates data
Error Analysis uses an API call (tested and approved) by ClickHouse which uses "Approximation".
The approximation algorithm (HyperLogLog) calculates whether some elements exist in an asset. We use three approximation algorithms based on the size of the data set.
- Small data set (If the limit is less than 9000) the data will be 100% accurate
- Medium data set: 0.05% margin of error
- Large data set (over 200,00 elements): average 0.1% margin of error
How to view the exact data
Open the error side panel to see the exact metrics for the error.
Go to the side panel to view the error details, and the KPI details based on 100% accurate data.
Example from a specific error from 'All Errors' tab:
Example from a specific error from 'Errors by Page Groups' tab:
You can continue your analysis with the Shortcuts to Session Replay, Impact Quantification, and Journey Analysis.