How to find patterns in the collected data?

Ask Me Anything Q&A Cxlama

Topic summary

The discussion centers on identifying patterns in user feedback, surveys, and questionnaires, with a focus on practical analytical methods.

Recommended Approaches:

  • Thematic analysis and codification are suggested as primary methods for breaking down both open-ended and choice-based responses into topics and themes
  • Use short user stories or tagging systems in databases/spreadsheets, then count occurrences
  • Apply severity ratings or weighting to interpret theme importance
  • A referenced blog post provides detailed guidance on codification techniques

Role of AI:

  • AI can assist in initial data digestion and provide directional insights
  • However, manual analysis is recommended first, as AI may hallucinate or miss nuanced opportunities
  • Once themes are manually identified, AI can help recognize those patterns across responses

Segmentation Strategy:

  • Analyze user metadata and traits (e.g., geographic location, user status) to spot behavioral patterns
  • Different user segments may exhibit distinct response characteristics
  • Target specific segments with tailored surveys or question variations using in-app tools
  • Example: returning users versus new users may demonstrate different behaviors and warrant separate research approaches
Summarized with AI on October 29. AI used: claude-sonnet-4-5-20250929.
1

What are the best practices in finding patterns in user feedback, survey, or questionnaires?

Also, what section of analytics reports do you suggest for understanding customers’ behaviours?

2

Hey @WingSpan

Thematic analysis and codification are my go to methods here. By doing so you can start to break down both open-ended and choice based answers into topics and themes.

I often use short user stories, or a form of tagging in a database or spreadsheet. Then I count those. Adding a severity or a weighting can often help interpret the importance of the themes and topics.

You can use AI to assist here, but I would advise against trusting it completely on this as it will hallucinate or miss opportunities you may be more qualified to spot. It can easily digest the information and give you a quick glimpse and some form of direction or idea, but I would keep it at just that. Once you have your main themes and topics from doing it yourself, you can then return to AI again to help identify those themes in the answers.

3

Like both ideas. They sound very practical.

Can you speak a bit more to how to do codification?

4

Thanks for laying out the tips! What caught my eye was segment.

How do we do that? In your experience, what are the easiest way to set up a workflow and get feedback from customers of individual segments?

5

Hey @WingSpan

This blog actually breaks it down quite well.

https://speero.com/post/codification-for-user-data-processing

6

I understood that using meta data or traits of the users from the existing responses is one way of spotting patterns.

For example all orders from US, UK, EU etc may exhibit different characteristics in their responses. For example using targeting tools in the survey, or in-app tools to send each segment different questions, or variations of the same surveys.

Or a returning user may do X, whilst a new user does Y.

You can always target those users separately with different surveys, or research types.