The Topics by net sentiment insight element shows the Net Sentiment Score of each topic, broken out by the values of another column. Use it to see how people feel about a topic in each segment, brand, or region, not just how often they mention it.
When to use it
Use this element when you want to answer questions like:
- Is Customer Service discussed more positively in Germany than in the USA?
- Which topics drive negative feedback for one brand but not for others?
- Where does sentiment on a topic differ significantly from the overall average?
If you want to know how often a topic is mentioned per segment, use the Topic Breakdown / Heatmap instead.
Key components
- Rows: Represent topic categories. Click a category label to drill down into its topics.
- Columns: Represent the values of the breakout column. Each column header shows the number of responses for that value (for example, Ryanair (n=2k)).
- Cell values: Show the net sentiment of the topic or category within that column value, from −100 to +100. A cell is empty when the topic has no mentions for that column value.
- Color scale: Five color bands from strongly negative to strongly positive: −100 to −60, −60 to −20, −20 to 20, 20 to 60, and 60 to 100. Strongly positive and strongly negative cells stand out at a glance.
How net sentiment is calculated
For each topic and breakout value, Caplena takes all mentions of the topic and calculates:
Net sentiment = % positive mentions – % negative mentions
Neutral mentions count toward the total, so they pull the score toward 0.
- +100: Every mention of the topic is positive.
- 0: Positive and negative mentions balance out, or most mentions are neutral.
- −100: Every mention of the topic is negative.
Hover over a cell to see:
- Net sentiment
- Mentions: The total number of mentions of the topic in that column value
- Positive, Neutral, and Negative mention counts
- Statistical significance: The p-value, if the difference is significant
Check the mention count before drawing conclusions. A net sentiment of −100 based on two mentions is far less reliable than −40 based on 500 mentions.
Statistically significant differences
Caplena tests whether a topic’s net sentiment in a column value differs from its net sentiment across all responses. Cells with a significant difference (p ≤ 0.05) are marked with an asterisk (*).
Heatmap and bar chart views
Switch between two views using the widget dropdown:
- Heatmap: A grid of topics and column values colored by net sentiment. Best for scanning many topics and segments at once.
- Bar chart: Bars showing net sentiment per topic and column value. Best for a direct side-by-side comparison of a few segments.
Choose which topics to show
Use the displayed topics settings to control which rows appear:
- Top N topics or Top N categories: Shows the topics or categories with the most mentions across all column values.
- Topic selection or Category selection: Shows only the topics or categories you pick.
Topics are ranked by mention volume, not by net sentiment. This keeps rarely mentioned topics with extreme scores from crowding out the topics that matter most.
Drill down into responses
Click a cell to open the responses behind it. Caplena filters the rows to the selected topic or category and the selected breakout column value.
Compared with Topic Breakdown / Heatmap
Use both elements side by side. A topic that is mentioned often and has a low net sentiment is a strong candidate for action.
Last modified on October 12, 2026
