Skip to main content

Retrieving User Feedback

This guide explains how to access and analyse user feedback and satisfaction survey data collected by Tellr.

Overview​

Tellr collects two types of user feedback:

  • Survey responses — Star ratings (1-5), time-saved estimates, and NPS scores (0-10) from the post-generation satisfaction survey.
  • Feedback conversations — Structured summaries from the AI feedback widget, categorised by type (Bug Report, Feature Request, UX Issue, Performance, Content Quality, Other) and severity (Low, Medium, High).

Both are stored anonymously in the database. The primary way to view this data is the Feedback Dashboard on the Admin page. For programmatic access, a reporting API and direct database queries are also available.

Prerequisites​

  • Access to the Tellr application

The Feedback Dashboard​

The Admin page includes a Feedback Dashboard that displays aggregated survey metrics and an AI-generated summary of recent feedback.

Step 01: Open the Admin Page​

Navigate to /admin. The Feedback tab is selected by default.

Admin page with Feedback tab

Step 02: Review Summary Metrics​

The top of the dashboard shows six summary cards covering the entire reporting period:

Summary metric cards

CardDescription
Distinct UsersNumber of unique users who submitted surveys
Total SessionsTotal slide generation sessions recorded
Survey ResponsesTotal number of survey submissions
Avg Star RatingOverall satisfaction out of 5
Avg NPS ScoreNet Promoter Score average (0-10)
Total Time SavedCumulative self-reported time savings

Step 03: View Weekly Survey Stats​

The Weekly Survey Stats table breaks down survey responses by week, showing average star rating, average NPS, and time saved per week. Use the Weeks dropdown to adjust the time window (4, 8, 12, 26, or 52 weeks).

Weekly Survey Stats table

Step 04: Read the AI Feedback Summary​

The AI Feedback Summary section uses an LLM to analyse recent feedback conversations and produce a narrative summary. It includes:

  • A natural language summary of themes and common issues
  • Top Themes — The most frequently mentioned topics
  • Category Breakdown — Pill-style tags showing the count of feedback items per category

Use the Weeks dropdown to adjust the summary period (2, 4, 8, or 12 weeks).

AI Feedback Summary

Understanding NPS Scores​

Net Promoter Score classifies respondents into three groups:

ScoreClassificationMeaning
9-10PromoterEnthusiastic, likely to recommend
7-8PassiveSatisfied but not enthusiastic
0-6DetractorUnhappy, unlikely to recommend

NPS = % Promoters minus % Detractors. A positive NPS indicates more promoters than detractors.

Reporting API​

For programmatic access or automated reporting, the same data is available via REST endpoints.

Weekly Stats​

GET /api/feedback/report/stats?weeks=12

Returns weekly aggregated metrics and overall totals. The weeks parameter defaults to 12.

{
  "weeks": [
    {
      "week_start": "2026-02-09",
      "week_end": "2026-02-15",
      "responses": 15,
      "avg_star_rating": 4.3,
      "avg_nps_score": 8.1,
      "total_time_saved_minutes": 960,
      "time_saved_display": "16h 0m"
    }
  ],
  "totals": {
    "total_responses": 87,
    "avg_star_rating": 4.2,
    "avg_nps_score": 7.8,
    "total_time_saved_minutes": 5400,
    "time_saved_display": "90h 0m"
  },
  "usage": {
    "total_sessions": 142,
    "distinct_users": 23
  }
}

AI Summary​

GET /api/feedback/report/summary?weeks=4

Returns an LLM-generated narrative summary. The weeks parameter defaults to 4.

{
  "period": "Last 4 weeks",
  "feedback_count": 12,
  "summary": "Over the past 4 weeks, users submitted 12 feedback items...",
  "top_themes": ["Template customisation", "Chart formatting"],
  "category_breakdown": {"Feature Request": 5, "Bug Report": 3, "UX Issue": 2}
}

Direct Database Queries​

For advanced analysis, query the database tables directly.

Survey Responses​

SELECT star_rating, time_saved_minutes, nps_score, created_at
FROM survey_responses
ORDER BY created_at DESC
LIMIT 50;

Feedback Conversations​

SELECT category, severity, summary, created_at
FROM feedback_conversations
ORDER BY created_at DESC
LIMIT 50;

Example: Monthly NPS Breakdown​

SELECT
  DATE_TRUNC('month', created_at) AS month,
  COUNT(*) AS responses,
  ROUND(AVG(nps_score), 1) AS avg_nps,
  COUNT(CASE WHEN nps_score >= 9 THEN 1 END) AS promoters,
  COUNT(CASE WHEN nps_score BETWEEN 7 AND 8 THEN 1 END) AS passives,
  COUNT(CASE WHEN nps_score <= 6 THEN 1 END) AS detractors
FROM survey_responses
WHERE nps_score IS NOT NULL
GROUP BY DATE_TRUNC('month', created_at)
ORDER BY month DESC;

Tips​

  • Start with the dashboard — The Admin page gives you an at-a-glance view without needing API tools
  • Adjust time windows — Use the Weeks dropdowns to focus on recent trends or longer-term patterns
  • Automate weekly reports — Call GET /api/feedback/report/stats from a scheduled job and post results to Slack or email
  • Time savings ROI — Multiply total time saved by an average hourly rate to quantify ROI in monetary terms
  • Combine stats and summary — Use the stats endpoint for numbers and the summary endpoint for qualitative insights in management reports