Process VoC data for reliable insights

Business benefits

Transform data into information and knowledge.

Create a spreadsheet with its first column titled with your question.

If you have VoC data from multiple questions, start a new sheet for each.

Add VoC quotes from customer surveys, interviews, and user tests to the question column of your spreadsheet.

List common themes that you see in the VoC quotes.

Group your themes into a few categories. Add these to column titles.

For example, Price is too high, Quality is good, Improves my day.

Score each response’s relevance to each category with a 0 for irrelevant or 1 for relevant.

Tally up the score for each category column and note the highest-scoring categories.

Use coding tools when the breadth of responses becomes too complex for manual processing.

Popular natural language processing tools include:

  • Amazon Comprehend, a DYI NLP service with integrated API.
  • Textalyser, a free topic modeling option for quick responses.
  • Sprig, a survey tool that also offers NLP.
  • Chattermill and Luminoso, enterprise-level NLP tools for complex surveys.

Develop simple statements that summarize the highest-scoring categories of themes.

For example, if the highest scoring categories are Price is too high and Quality is good, a statement might be Customers don’t like the high price, but find the quality worthwhile.

Use your statements as hypotheses for building value propositions and running marketing experiments to improve your brand’s digital experience.

Last edited by @hesh_fekry 2023-11-14T10:08:50Z

Could you clarify what to look for, and how to categorize information in this process?

I am using the Message Mining Data-Entry Form from the Product messaging course.
It provides further details, like what to look for and how to categorize your entries. But I find that some categories are missing in it, like frustration.

I’m using reviews from competitors as I haven’t started my business yet.
I have found a wealth of reviews on Trustpilot, and this process is really eye opening.
But these are reviews not surveys. I’m not sure how to categorize negative reviews elements as there are not categories for pain points.

I do find things that that fit the deal-breaker feature… but I’m not sure about frustration.
Also, I’m not getting much info about the initial pain points but rather desired outcome.

I do see patterns emerge though, and I’m getting insights about what’s important to people. I even see how I will have to structure my team and service delivery.

I still would love some clarification on how to categorize the data.

1 Like

Thank you Laura for the comment. I will tag one of our instructors who may have some suggestions about this: @ben2 ( Instructor for Voice of Customer Data )

1 Like

Thanks. I haven’t received any replies yet, but I’ll just add the category “pain points and frustrations” to my list of categories… and I think I’ll be good.