Plan requirement
| Subscription | Any plan |
| Also required | Quality Assurance or Workforce Engagement Management |
| Access | Admin |
Filtering and exporting the quality data, and the filters that make a number mean something different.
Work with it
- Open the dashboard for the workspace.
- Set the period. The This week, This month and This year ranges count today as one of the days, so a Monday morning figure rests on one day.
- Filter by group, channel, reviewer or reviewee, and sort the table by date, score, pass rate, review time, whether the reviewee has seen it, disputes, assignment or comments.
- Download as CSV or PDF if somebody needs it elsewhere. The download follows whatever filters are set, and it still contains data you have hidden from the visualisation.
The headline figure is IQS, the internal quality score: the average of all manual review scores given or received over the period, as a percentage. Automated reviews are not in it; they have their own figure, AQS, which is the plain average of the AutoQA scores across every conversation it evaluated.
The AutoQA side refreshes on its own once an hour, so a score that has not appeared yet is usually a matter of waiting rather than a problem to chase.
Use a long enough period
A month of reviews for a small team is a handful of conversations. A quarter gives you something stable enough to act on.
Filter by reviewer occasionally
If one reviewer scores consistently lower than another on the same scorecard, that is a calibration problem rather than a quality difference.
It is the check nobody thinks to run, and it explains a surprising number of odd results.
Filter by channel
Email and chat are judged against the same scorecard and behave differently. A combined score can hide one channel performing much worse.
Do not compare across workspaces
Different scorecards, different reviewers, different standards. Two workspace averages side by side is the most common misreading of this data.
Exporting
Fine for analysis elsewhere. It contains individual scores and comments about named colleagues, so decide where the file lives and who opens it before sending it.
Note the period on anything you share
Quality numbers get quoted months later. A figure without its period and its scorecard version will be compared with something incomparable.
Comments
0 comments
Article is closed for comments.