Scores

Once you understand how applications are designed and constructed, you will understand what the application is measuring, how the application is measuring, and what the application metrics mean.

Note that in some cases, a "low" score may indicate a positive outcome, and a "high" score may indicate a negative outcome. For example, if an application measures the use of impolite language by Agents, a "high" score could indicate that some Agents are speaking rudely to callers.

There are numerous ways to view application scores in a variety of contexts:

  1. Using the application breadcrumb links on the dashboard icon Dashboard:

    1. The application dashboard shows two graphs:

      1. The Overview shows coverage and hit/miss scores for top-level application categories for all days in the specified time period. Each day has one chart column.

      2. The Daily Chart shows a detailed breakdown of the selected day's scores. To change days, click a column in the Overview.

    2. Below the graphs, the application dashboard shows three views:

      1. Agents — View scores for individual agents.

      2. Summary — View and compare daily summary scores.

      3. Files — Define search terms and other parameters to review specific transcripts.

  2. The Application Scores tab on a transcript's File Details page.

  3. The /appmatches endpoint.

Best practices

When defining the analytic constraints within an application, you are advised to define the categories and filters so that high scores indicate either positive or negative aspects of the transcript, and not both. This best practice will make it easier to understand the resulting scores quickly. For example, if you assign scores that indicate both positive and negative aspects of a call, then more attention is required to determine if "high scoring" agents are functioning well or poorly.

If leaf-level categories are not resulting in the scores you expect, review the filters of the "parent" category at the level above the leaf-level. It is most likely that one or more filters are preventing the transcript from being passed down to the leaf level.