Prodsight Topics can help you quickly make sense of thousands of customer conversations. However, the quality of your insights greatly depends on how well your topics are structured. Please read this guide to learn more about how to define good topics.
What makes a good topic?
Attributes of a good topic:
- The topic name reflects the mention matches well
- Topic keyword list is robust and extensive
- The topic is meaningful and important to your organization or team
✅Example of a good topic
- Topic name: "Technical Issues"
- Keywords: "bug", "error", "crash", "doesn't work", "broken", "blank screen"
This topic is good because its keyword list is extensive and captures various ways users might be describing technical issues. The name describes the contents of the topic well. Finally, tracking product issues is generally something that might be important for your organization to keep an eye on.
❌Example of a bad topic
- Topic name: "Feedback"
- Keywords: "billing", "pricing", "help", "issue", "agent", "playlists"
This topic has a broad keyword list capturing a variety of customer topics which means that mentions will be very different from each other. This will make isolating specific trends or issues difficult. The name of the topic is broad and not descriptive of the contents and is unlikely to help your team or organization much.
How to come up with good topics?
A good starting point is to break down your customer conversations into broad categories. For example, if you run a software product, your categories could include:
- Bugs (bug, error, crash, issue, fault, etc.)
- Usability Issues (UX, UI, intuitive, slow, interface, font, etc.)
- New Feature Requests (feature idea, feature suggestion, add feature, etc.)
- User Education (how do I, way to, etc.)
- Finance (billing, price, cost, card, payment, etc.)
- Generic Negatives (hate, bad, poor, unhappy, etc.)
- Generic Positives (great, awesome, amazing, love, etc.)
Another useful entry point is to think of all major product areas or features in your offering and define a topic for each. For example, for a music streaming service, these might include:
- Player (play, stream, stop, pause, skip, song, etc.)
- Playlists (playlist, library, daily mix, discover weekly, etc.)
- Artists (artist, performer, bio, album, discography, etc.)
- Apps (app, iOS, Android, app store, play store, download, etc.)
Your team will undoubtedly be aware of some recurring issues reported by your customers. Get together with few support agents as well as people from product, marketing and sales teams and brainstorm a list of known issues.
These might include:
- Churn indicators such as cancellations or returns
- Existing bugs
- Known usability issues
- Frequently asked questions
- Known concerns or sales objections
- Mentions of competitors
By creating a topic for each known issue you will be able to easily quantify the impact in terms of volume and sentiment and prioritize improvements based on data rather than gut.
We hope this guide will give you some ideas to get started. If you'd like more hands-on help or advice on how to structure topics for your particular case, do not hesitate to contact us via live chat or email.