Help Guides
Surveys
Reporting
Getting Responses
Managing Responses
Topics and sub-topics are how CustomerSure helps you understand what customers are talking about, how they feel about it, and where to focus improvement effort first.
They sit at the heart of driver analysis, heatmaps, league tables, and AI summaries.
What are topics?
Topics are things your organisation does that customers can express a positive, neutral, or negative sentiment about.
Examples of good topics:
- Speed
- Communication
- Cleanliness
- Ease of use
- Value for money
A topic should always be something you could realistically take action on. If a team can’t own it and improve it, it’s probably not a good topic.
Topics are not categories for ‘filing’ feedback — that’s what labels are for. Topics always have sentiment attached.
What are sub-topics?
Sub-topics sit underneath topics and represent the specific things you can actually work on.
For example:
-
Topic: Communication
- Sub-topic: Kept me informed
- Sub-topic: Explained next steps clearly
- Sub-topic: Easy to contact
-
Topic: Cleanliness
- Sub-topic: Rooms
- Sub-topic: Bathrooms
- Sub-topic: Communal areas
Topics are broad, and sub-topics as the specific, actionable things inside them.
Sentiment and topics
For any topic or sub-topic, customers can express:
- Positive
- Neutral
- Negative
When you’re looking at a single topic, sentiment is never mixed. You’ll only see ‘mixed’ sentiment if you’re looking at a piece of feedback which refers to multiple topics.
Topic sentiment comes from two sources:
1) Rating questions
Rating questions can feed into one or more topics.
For example, a question like:
“How would you rate the speed of our service?”
can be linked to the Speed of service sub-topic.
Positive scores contribute positive sentiment, and negative scores contribute negative sentiment.
See rating questions for full details on how questions work.
2) Comment box analysis (AI)
If AI analysis is enabled, CustomerSure will scan customer comments and:
- Detect which topics and sub-topics are being discussed
- Assign sentiment based on the language used
This allows you to keep surveys short (often a single comment box) while still getting detailed, structured insight.
Analysing topics and sub-topics
For every topic and sub-topic you can see:
- Sentiment breakdown (positive / neutral / negative)
- The customer comments that contributed to that sentiment
This makes it easy to move from a headline number to real customer language, without guesswork.
AI summaries and action plans
For sub-topics, CustomerSure can generate:
- An AI-written summary of what customers are saying. Useful for stakeholders who don’t want to read hundreds of comments.
- An improvement action plan: A structured starting point for fixing the issue, based on the feedback received.
These are designed to accelerate decision-making, not replace human judgement.
League tables
You can rank topics or sub-topics in league tables, to help spot, for example:
- Best → worst topics overall
- Weakest sub-topics within a specific journey or business area
This helps answer questions like “What should we work on first?”
Heatmaps
Heatmaps give more detail than league tables. Rather than just looking at a list of ranked topics, you can segment any topic by your reporting segments: staff, teams, areas, organisational groups, and find the real hotspots in your organisation.
This makes it easy to spot patterns such as:
- A strong topic overall, but weak in one location
- A recurring issue affecting a specific team or region
Driver analysis
Driver analysis answers the question:
Which topics matter most to our top-line metrics?
CustomerSure can show you which topics and sub-topics strongly correlate with metrics like NPS, CSAT, or CES.
This allows you to prioritise improvements based on impact, not volume or gut feel:
- A topic mentioned less often may be a stronger driver of churn
- A frequently mentioned topic may have little effect on overall satisfaction
Configuring topics and sub-topics
Topics are configured here:
Company Settings → Feedback Structure → Topics
Key points:
- Topics are mostly containers. They exist to organise sub-topics.
- Sub-topics do the real work.
For each sub-topic you must provide a description.
This description is used by the AI to decide:
- When the sub-topic applies
- What sentiment should be assigned
A good description is important.
Good vs poor sub-topic descriptions
Good description
“Mentions of staff clearly explaining what will happen next, timelines, or follow-up steps.”
Poor description
“Communication issues.”
Good description
“Mentions of prices being unclear, unexpected charges, fees not explained in advance, or confusion about what was included.”
Poor description
“Pricing problems.”
Good description
“Mentions of staff attitude, friendliness, politeness, empathy, or how the customer was treated by an individual.”
Poor description
“Staff issues.”
Topics vs labels
| Topics | Labels |
|---|---|
| Always have sentiment | Never have sentiment |
| Represent things customers feel about | Represent categories of feedback |
| Used in drivers, heatmaps, league tables | Used for filtering and workflow |
| Feed strategic improvement | Feed operational handling |
Both are important, but they serve different purposes.
Next steps
Getting topic and sub-topic definitions right is one of the foundations of a VoC programme that delivers ROI.
If your sub-topics are vague, generic, or poorly defined, everything downstream suffers: Analysis becomes noisy, and teams lose confidence because the insight doesn’t clearly point to action. Every business has suffered from VoC reporting for its own sake, rather than being used as a tool for change.
When topics and sub-topics are well designed, the opposite happens. Feedback flows cleanly into priorities, priorities translate into ownership, and improvements can be tied back to movement in NPS, CSAT or CES.
This is where CustomerSure is different. We don’t just give you a platform and leave you to guess. We work closely with customers to design topic structures that reflect how your organisation actually works. If you need help, get in touch
