Sentiment labels provide a quick and easy indication of whether a comment is positive, neutral, or negative, displayed right next to the comment text on the Comments overview page. Comment sentiment shows how users felt about content based on comments they left on news posts. It is calculated as the difference between the percentage of positive and negative comments. The calculated score ranges between -100 and +100.

Sentiment analysis is performed by processing and labeling all comments of posts. The processing is done via an in-house pipeline using a machine-learning model. This pipeline provides probabilities for each comment being labeled positive, neutral, or negative. The model supports multiple languages. Since it runs fully in-house, Staffbase does not require any third-party processor to perform the analysis.

Filtering by Sentiment

You can filter the comments on the overview page to only display comments with specific sentiments.

1. In the Studio, navigate to Content > System > Comments.
   The Comments overview opens.
2. Click Sentiment to open the filter dropdown.
   
   
3. Select one or more sentiments to filter the comments.
4. Click Apply.
   The filtered comments display.
   
   

You have filtered the comments by sentiment. You can remove the filter selection at any time to display all the comments again.

Editing the Sentiment of a Comment

You can manually edit a comment if it has an incorrect sentiment. The changes are reflected in the Comment Sentiment section of the Smart Insights dashboard.

1. In the Studio, navigate to Content > System > Comments.
2. Navigate to the comment for which you want to adjust the sentiment.
3. Click the three dots to open the action menu.
4. Click Edit Sentiment.
   
   
5. From the dropdown, select the new sentiment.
6. Click Save.

You have adjusted the sentiment for a comment.