EN 
FR 
Customer issue
Customer story: Automated processing of user feedback for a media company
This company in the educational media received a steady stream of customer feedback every week:
-
Opinions on Trustpilot,
-
Answers to satisfaction forms,
-
Reasons for unsubscribe expressed by users.
Over time, these feedbacks accumulated... until they formed a veritable mountain of comments.
The problem?
These data were scattered across several tools, unstructuredand therefore almost impossible to exploit clearly for continuous improvement.
The team wanted to better understand expectations, frustrations and strengths perceived by their users. But without a dedicated data team, or a centralized solution, it was very difficult to make decisions based on this feedback.
Context
In a fast-growing B2C company, customer feedback is essential for product development.
But when the information is fragmented between several sources (opinion platforms, internal tools, emails...), it becomes complicated to spot trends, or even to know what to prioritize.
Work carried out by ActivDev
Challenges
-
Too many different sources → Trustpilot, internal forms, cancellation pages, etc.
-
No clear structure → hard to know what users are talking about (price? bugs? content?)
-
No sentiment analysis → impossible to know what is positive or negative overall
-
No tracking over time → no overview of the evolution of returns
Tools used
We have implemented an automated, easy-to-maintain, code-free solution that combines several powerful tools:
-
n8n Automated data collection and processing
-
OpenAI to analyze the your feedback (positive, neutral, negative), allocate a categoryand extract keywords
-
Airtable to centralize enriched data and create a visual dashboard
-
Native APIs to connect Trustpilot, internal tools and Airtable
What we have put in place
1. Automatic collection, every day
Every day, an n8n scenario automatically retrieves new returns via API :
-
New Trustpilot reviews
-
Responses to satisfaction surveys
-
Reasons for cancellation given by users
2. Intelligent analysis and enrichment
Each return is processed automatically:
-
Sentiment positive, neutral or negative
-
Category interface, pricing, bugs, content, support, etc.
-
Keywords identification of recurring terms (via AI)
3. Storage in Airtable + dashboard
The data is stored in a clear Airtable database that can be consulted by the whole team.
4. Weekly visual report
Each week, a dashboard highlights :
-
Total number of returns
-
Positive vs. negative feelings
-
Trends in most frequently cited categories
-
A "Quick Wins" section for easy-to-implement improvements
- A section analyzing background problems, recurring keywords and their context, with a view to improving the category concerned.
Results
Clear vision of priorities the most frequently mentioned problems are identified at a glance
Improved responsiveness many of the irritants detected were quickly rectified
Fact-based decisionsnot on impressions
Time saved no need to read each notice manually or summarize by hand
Results
Automation of customer returns from 0 to 100%