Exporting Data - Tellet Help Center

Download conversation data as Excel, CSV, or TXT files for external analysis, backup, or sharing. Excel (.xlsx) is the default and opens cleanly with no setup.

Export Locations

From Transcripts Table

Export multiple conversations: in the Transcripts Tab, click Export (top-right area), choose what to export and a file format, then click Export to download.

This includes all conversations in the current view (Completed or Incomplete), narrowed to any active filters, for the export type you choose.

Media Download (Single Conversation)

From a participant's transcript you can also download a ZIP archive containing all the media that participant shared (video, audio, and images).

From Single Conversation

Export one conversation: open it in the Transcripts Tab, click Export, select a format (Excel, CSV, or TXT), then click Export to download. The file contains the full transcript with all messages, all metadata for that participant, and a timestamp for each message.

The export always contains every available language version of each message, regardless of which translation you have selected in the transcript view.

From Overview or Questions Tab

Export an analysis text file: in the Overview or Questions Tab, click "Download Analysis Text File" (top-right area, next to the filter selector). It downloads automatically as TXT.

Exported data includes:

Project statistics (completed interviews, started interviews, completion rate, average/median time)
Overall summary
Overall themes and opportunities
Overall interesting quotes
Per-question insights (summary, themes, opportunities, quotes)
Categories with counts and percentages for each question

Note: If a filter is applied, the export will only include analysis data for conversations matching that filter. Hover over the button or info icon to see which filter is applied.

Export Types

When exporting from the Transcripts Table you choose one of three layouts. The names describe the row layout (one row per respondent vs. one row per message):

Answers in columns (one row per respondent) — one column per question. Best for analysis in Excel, SPSS, or coding tools. This is the default.
Full conversation (one row per message) — every interviewer and respondent turn on its own row. Best for reading the whole conversation.
Summary & categories (one row per respondent) — summary, categories, and metadata per respondent, without transcripts.

Want one column per question? Choose Answers in columns — the Full conversation layout puts every message on its own row instead.

Answers in Columns

One row per respondent, with each question in its own column. This is the best layout for analysis and coding.

Included data:

Conversation ID, participant number, start time, duration, and summary
One column per question containing that respondent's answer
A translated column per question for each available language (when translations exist)
Category labels assigned per question
All metadata fields
Presentation order — for every question, the position (1st, 2nd, 3rd…) at which this respondent actually saw it. This captures question and question-group randomisation, so you can tell the order each respondent experienced.
Stimulus columns — for questions that show media (images, video, audio), the identifier and label of the stimulus shown, in the order shown. This tells you which randomised concept a respondent saw.

A second row directly under the header repeats the full text of each question, so the file is readable on its own.

Use cases:

Concept tests and monadic/sequential-monadic designs
Order-effect and randomisation analysis
Statistical analysis in SPSS, R, or Python
Coding in qualitative tools that expect one row per respondent

Summary & Categories

Summary data for each conversation.

Included data:

Conversation ID
Participant number
AI-generated summary of the conversation
Start time (when the participant sent their first response)
Duration (HH:MM:SS format)
Category labels assigned per question (one column per question)
All metadata fields (age, gender, location, custom screening data)

Excluded data:

Message content
Full transcripts
Individual message-by-message responses

Use cases:

Demographic analysis
Completion rate analysis
Timeline tracking
Quantitative analysis in spreadsheets
High-level data sharing

Full Conversation

Full transcripts with all messages.

Included data:

All Summary & categories data, plus:
Complete message-by-message transcripts (one row per message)
A message identifier and timestamp for each message
Message sender (Interviewer or Participant)
Translated message columns (one per language) when translations exist
Question ID — the identifier of the question each message relates to, so you can join answers to questions in SPSS, NVivo, or Atlas.ti
Presentation order — the position at which this respondent saw the related question (captures randomisation)
Answer source — whether a respondent answered by voice (transcribed audio) or by typing
Media columns — the count, file identifiers, and types of any attachments on a message, in the order shown

File size: Larger (varies based on conversation length)

Use cases:

Qualitative analysis software (NVivo, Atlas.ti)
Content analysis
Quote extraction
Manual coding
Complete data backup
Research documentation

Analysis Report

AI-generated analysis and insights for the project.

Included data:

Project title and filter information (if applied)
Overall statistics:
- Completed interviews count
- Started interviews count
- Completion rate percentage
- Average time to completion
- Median time to completion
Overall summary (AI-generated synthesis)
Overall themes (key themes identified across all conversations)
Overall opportunities (actionable insights)
Overall interesting quotes (notable participant quotes)
Per-question analysis:
- Question text and index
- Summary for each question
- Themes specific to each question
- Opportunities for each question
- Quotes relevant to each question
- Categories with labels, descriptions, counts, and percentages

File format: TXT only

Use cases:

Sharing analysis insights with stakeholders
Creating presentations and reports
Archiving AI-generated insights
Comparing analysis across different filters
Quick reference for key findings

File Formats

Excel (.xlsx)

Native spreadsheet format and the default choice. Recommended for most users.

Advantages:

Opens cleanly in Excel on a double-click — no import wizard, no separator settings, and no garbled accented or non-Latin characters, on any computer or language setting
Each export includes a Data Dictionary tab that explains every column (what it means and which version of the export introduced it), so the new presentation-order, answer-source, and stimulus columns are self-explanatory
Right-to-left languages (such as Arabic and Hebrew) display in the correct direction, cell by cell, even in a study that mixes languages

Use Excel for:

Everyday analysis and sharing
Opening reliably across different computers and regional settings
Studies in multiple languages, including right-to-left scripts

CSV (Comma-Separated Values)

Structured spreadsheet format. Choose this when a downstream tool (pandas, R, a BI tool, or a data pipeline) expects plain CSV.

File structure:

conversation_id,respondent_index,start_time,duration,age,gender
abc123,1,2025-01-15T14:30:00Z,00:20:00,25,Male
def456,2,2025-01-15T15:00:00Z,00:17:50,30,Female

Advantages:

Opens in Excel, Google Sheets, Numbers
Supports pivot tables and charts
Compatible with statistical software (SPSS, R, Python)
Automatic column alignment

Use CSV for:

Quantitative analysis
Statistical analysis
Charts and visualisations
Spreadsheet work

TXT (Plain Text)

Simple text format with tab or comma separation.

File structure:

conversation_id    respondent_index    start_time     duration
abc123            1                   2025-01-15     00:20:00
def456            2                   2025-01-15     00:17:50

Advantages:

Opens in any text editor
Lightweight
No formatting complications
Smaller file size

Use TXT for:

Manual review
Plain text archival
Import to qualitative analysis software
Minimal file size requirements

Export Process

Click Export (Transcripts table or single conversation) to open the export modal
Choose what to export (Answers in columns, Full conversation, or Summary & categories) — skipped for a single conversation
Select a format (Excel, CSV, or TXT) — Excel is selected by default
Click Export
The file is prepared and downloads to your browser's Downloads folder

Export Data Structure

Summary & Categories Columns

Standard columns:

conversation_id - Unique identifier
respondent_index - Participant number (1, 2, 3...)
summary - AI-generated summary of the conversation
start_time - ISO timestamp of when the participant sent their first response
duration - Time to complete in HH:MM:SS format (e.g., 00:19:55)
categories for Q1, categories for Q2, ... - Category labels assigned to each conversation per question

Metadata columns:

All metadata fields collected during the interview (e.g., participant_id, source, age, gender, language, custom screening answers)
Columns are dynamic and vary per project based on what metadata was captured

Full Conversation Columns

Each row represents one message. Conversation-level fields appear on the first message row only.

Columns:

conversation_id - Unique identifier
respondent_index - Participant number (1, 2, 3...)
start_time - ISO timestamp of first response
duration - Time to complete (HH:MM:SS format)
summary - AI-generated summary (first row only)
_id - Unique message identifier
role - "Interviewer" or "Participant"
created_at - Message timestamp
original_message - Message text content
translated_message_{language} - Translated content (one column per language if translations exist)
Metadata fields (first row only): all metadata captured during the interview (e.g., participant_id, source, age, gender)
question_id - Identifier of the question this message relates to (join key for SPSS/NVivo/Atlas.ti)
presented_order - The position at which this respondent saw the related question (reflects randomisation; blank for older data)
answer_source - voice (transcribed audio) or typed; blank for interviewer turns
media_count, media_files, media_types - Attachments on the message and their identifiers/types, in the order shown

Newer columns (question ID, presentation order, answer source, media) are added at the far right, so any existing column keeps its position and scripts keyed on the old layout keep working.

Example row structure:

conversation_id,respondent_index,start_time,duration,summary,_id,role,created_at,original_message,age
abc123,1,2025-01-15T14:30:00Z,00:12:30,AI summary text,msg001,Interviewer,2025-01-15T14:30:05Z,"Welcome!",25
abc123,,,,,msg002,Participant,2025-01-15T14:30:20Z,"Hello there",
abc123,,,,,msg003,Interviewer,2025-01-15T14:30:30Z,"What is your goal?",

Answers in Columns Export Columns

One row per respondent. The first row is the header; the second row repeats each question's full text as a label.

Columns, left to right:

conversation_id, respondent_index, start_time, duration, summary
One column per question (Q1, Q2, …) holding that respondent's answer, plus a translated column per question for each language when translations exist
categories for Q1, categories for Q2, … - Category labels per question
All metadata fields
Q1_presented_order, Q2_presented_order, … - The position (1st, 2nd…) at which this respondent saw each question; reflects question/group randomisation, blank for older data
Q{n}_stimulus_id, Q{n}_stimulus_label - For questions that show media, the identifier and label of the stimulus shown, in the order shown (only present for questions with media)

The presentation-order and stimulus columns are added at the far right so existing columns keep their positions.

Data Dictionary (Excel only)

Excel exports include a second tab named Data Dictionary. It lists every column in the file with a short description and the export version that introduced it, plus the overall export version. Use it as a quick reference for what each column means — especially the presentation-order, answer-source, and stimulus columns.

Single Conversation TXT Format

Formatted as a readable conversation, with a header block, an optional summary and category breakdown, metadata, then the full transcript grouped by speaker:

================================================================================
TRANSCRIPT #1
================================================================================
Conversation ID: abc123
Respondent ID: participant-abc
Respondent Index: 1
Date: 2025-01-15 14:30:00
Duration: 20 minutes

SUMMARY:
    AI-generated summary of the conversation.

CATEGORIES:
  Q1: Convenience

METADATA:
  age: 25
  gender: Male
  location: New York
================================================================================

CONVERSATION TRANSCRIPT:

INTERVIEWER (2025-01-15 14:30:05):
    [P1]     Welcome! Let's begin. What is your main goal?

RESPONDENT (2025-01-15 14:31:00):
    [P1]     To save time

Working with Exports

Opening in Excel

Double-click CSV file, or
Open Excel → File → Open → Select CSV

CSV exports are built to open correctly in Excel with no extra steps. The data lands in separate columns, and accented or non-Latin characters (for example Turkish or Japanese) display properly straight away — no import wizard needed.

The notes below cover rare cases where a file still needs a small adjustment, usually depending on your Excel version or system settings.

Dates display incorrectly:

Select date column → Format → Number → Date

Numbers not calculating:

Select column → Data → Text to Columns → Delimited → Comma

Special characters show as symbols:

This is rare. If it happens, re-open or re-import the file and set the
encoding to UTF-8.

Opening in Google Sheets

Open Google Sheets
File → Import
Upload tab → Select CSV file
Choose "Comma" as separator
Click Import

Analysis in Excel

Completion rate by demographic:

Create pivot table
Rows: demographic field (age, gender, etc.)
Columns: status
Values: count of conversations

Average duration by segment:

Create pivot table
Rows: segment variable
Values: average of duration

Keyword search in responses:

Use Find feature (Ctrl+F / Cmd+F)
Search the original_message column
Filter matching rows

Qualitative Analysis Software

NVivo:

Export as TXT
Import as dataset
Code transcripts
Run queries

Atlas.ti:

Export as TXT or CSV
Import as primary documents
Create codes
Analyse with quotation tools

Dedoose:

Export as CSV
Import as excerpts
Add descriptors from metadata
Code and analyse

MAXQDA:

Export as TXT
Import as documents
Use auto-coding features
Run text search

File Management

Naming Convention

Use descriptive filenames with dates:

Format: ProjectName-ExportType-YYYY-MM-DD.csv

Examples:

UserResearch-SummaryCategories-2025-01-15.xlsx
CheckoutStudy-FullConversation-2025-02-01.xlsx
MarketResearch-Age25to34-2025-01-20.csv

Include:

Project name/identifier
Export type (Answers in columns / Full conversation / Summary)
Date (YYYY-MM-DD format)
Filter criteria if applicable (Age25to34, Completed, etc.)

Folder Structure

Recommended organisation:

Project-Name/
  ├── Raw-Data/
  │   ├── 2025-01-15-SummaryCategories.xlsx
  │   ├── 2025-01-22-FullConversation.xlsx
  │   └── 2025-02-01-Final.csv
  ├── Analysis/
  │   └── Coded-Data.xlsx
  └── Reports/
      └── Final-Report.pdf

Export Log

Track exports in a spreadsheet:

Columns:

Date of export
Export type (Answers in columns / Full conversation / Summary)
Number of conversations
Filters applied
Purpose/milestone
File location

Example:

Date       | Type     | Count | Filters    | Purpose       | Location
2025-01-15 | Summary  | 23    | None       | Week 1 backup | /Raw-Data/
2025-01-22 | Full conversation | 47 | Age 25-34 | Segment data | /Analysis/
2025-02-01 | Full conversation | 75 | None     | Final archive | /Raw-Data/

Backup Schedule

Export data at:

Every 25 completed conversations
Weekly during active collection
Before deleting conversations
Before major project changes
Project completion

Filtering and Exports

How Filters Affect Exports

Active filters determine exported conversations:

No filter: exports all conversations in current view
With filter: exports only matching conversations
Filter criteria not included in export file name (document manually)

Segment Export Strategy

Export full dataset (no filters) as master backup
Apply filter to segment A
Export with descriptive filename
Remove filter
Apply filter to segment B
Export with descriptive filename
Repeat for each segment
Compare in analysis tool

Multiple filter combinations:

Example: Age 25-34 AND Gender: Female AND Location: California
Filename: Project-Women25-34-CA-2025-01-15.csv

Troubleshooting

Export button missing or disabled

Possible causes:

No conversations exist
Permission issues
Browser cache holding a stale UI state

Solutions:

Verify completed conversations exist
Confirm you have the required role (Owner, Admin, Editor, or workspace-level export permission)
Refresh browser
Log out and log back in

Download failed or corrupt file

Possible causes:

Network interruption
Browser blocking download
File too large

Solutions:

Retry download
Use different browser (Chrome or Firefox)
Check internet connection
Export smaller segments using filters
Clear browser cache

File won't open in Excel

CSV exports normally open in Excel on a double-click. In the rare case one
doesn't, try:

Right-click → Open with → Excel
Excel: Data → Get Data → From Text/CSV
Set encoding to UTF-8
Try Google Sheets

Scrambled data or weird characters

Exports display accented and non-Latin characters correctly out of the box, so this is rare. If you do see scrambled characters (for example after editing and re-saving the file in another tool):

Re-open or re-import the file with UTF-8 encoding
In Google Sheets: import with UTF-8 selected
Open in a text editor to verify the data
Re-export a fresh copy from Tellet

Missing conversations

Possible causes:

Active filter
Wrong view (Completed vs Incomplete)
Very recent conversations may take a moment to appear

Solutions:

Remove all filters
Check both Completed and Incomplete tabs
Refresh page
Verify conversations exist in Transcripts table

Data Privacy and Security

Data Sensitivity

Exports may contain:

Personal information
Demographic data
Opinions and feedback
Sensitive topics

Best practices:

Store in secure locations
Don't email unencrypted files with sensitive data
Use password-protected folders
Delete old exports when no longer needed
Follow organisational data policies

Sharing Exports

Before sharing:

Review included data
Confirm recipient needs access
Consider anonymisation or redaction
Use secure file sharing
Set link expiration dates

Anonymisation:

Remove participant identifiers
Generalise locations or details
Redact personal information in quotes
Use aggregate data when possible

Use Cases

Early Project (Exploration)

Export Summary & categories to check demographic distribution
Verify data quality
Create backup

Mid-Project (Analysis)

Export Full conversation for qualitative coding
Create pivot tables for segment analysis
Generate custom visualisations
Combine with other datasets

End of Project (Documentation)

Export complete dataset for archival
Share with external researchers
Meet data retention requirements
Final backup

Download conversation data as Excel, CSV, or TXT files for external analysis, backup, or sharing. Excel (.xlsx) is the default and opens cleanly with no setup.

Export Locations

From Transcripts Table

Export multiple conversations: in the Transcripts Tab, click Export (top-right area), choose what to export and a file format, then click Export to download.

This includes all conversations in the current view (Completed or Incomplete), narrowed to any active filters, for the export type you choose.

Media Download (Single Conversation)

From a participant's transcript you can also download a ZIP archive containing all the media that participant shared (video, audio, and images).

From Single Conversation

The export always contains every available language version of each message, regardless of which translation you have selected in the transcript view.

From Overview or Questions Tab

Export an analysis text file: in the Overview or Questions Tab, click "Download Analysis Text File" (top-right area, next to the filter selector). It downloads automatically as TXT.

Exported data includes:

Project statistics (completed interviews, started interviews, completion rate, average/median time)
Overall summary
Overall themes and opportunities
Overall interesting quotes
Per-question insights (summary, themes, opportunities, quotes)
Categories with counts and percentages for each question

Note: If a filter is applied, the export will only include analysis data for conversations matching that filter. Hover over the button or info icon to see which filter is applied.

Export Types

When exporting from the Transcripts Table you choose one of three layouts. The names describe the row layout (one row per respondent vs. one row per message):

Answers in columns (one row per respondent) — one column per question. Best for analysis in Excel, SPSS, or coding tools. This is the default.
Full conversation (one row per message) — every interviewer and respondent turn on its own row. Best for reading the whole conversation.
Summary & categories (one row per respondent) — summary, categories, and metadata per respondent, without transcripts.

Want one column per question? Choose Answers in columns — the Full conversation layout puts every message on its own row instead.

Answers in Columns

One row per respondent, with each question in its own column. This is the best layout for analysis and coding.

Included data:

Conversation ID, participant number, start time, duration, and summary
One column per question containing that respondent's answer
A translated column per question for each available language (when translations exist)
Category labels assigned per question
All metadata fields
Presentation order — for every question, the position (1st, 2nd, 3rd…) at which this respondent actually saw it. This captures question and question-group randomisation, so you can tell the order each respondent experienced.
Stimulus columns — for questions that show media (images, video, audio), the identifier and label of the stimulus shown, in the order shown. This tells you which randomised concept a respondent saw.

A second row directly under the header repeats the full text of each question, so the file is readable on its own.

Use cases:

Concept tests and monadic/sequential-monadic designs
Order-effect and randomisation analysis
Statistical analysis in SPSS, R, or Python
Coding in qualitative tools that expect one row per respondent

Summary & Categories

Summary data for each conversation.

Included data:

Conversation ID
Participant number
AI-generated summary of the conversation
Start time (when the participant sent their first response)
Duration (HH:MM:SS format)
Category labels assigned per question (one column per question)
All metadata fields (age, gender, location, custom screening data)

Excluded data:

Message content
Full transcripts
Individual message-by-message responses

Use cases:

Demographic analysis
Completion rate analysis
Timeline tracking
Quantitative analysis in spreadsheets
High-level data sharing

Full Conversation

Full transcripts with all messages.

Included data:

All Summary & categories data, plus:
Complete message-by-message transcripts (one row per message)
A message identifier and timestamp for each message
Message sender (Interviewer or Participant)
Translated message columns (one per language) when translations exist
Question ID — the identifier of the question each message relates to, so you can join answers to questions in SPSS, NVivo, or Atlas.ti
Presentation order — the position at which this respondent saw the related question (captures randomisation)
Answer source — whether a respondent answered by voice (transcribed audio) or by typing
Media columns — the count, file identifiers, and types of any attachments on a message, in the order shown

File size: Larger (varies based on conversation length)

Use cases:

Qualitative analysis software (NVivo, Atlas.ti)
Content analysis
Quote extraction
Manual coding
Complete data backup
Research documentation

Analysis Report

AI-generated analysis and insights for the project.

Included data:

Project title and filter information (if applied)
Overall statistics:
- Completed interviews count
- Started interviews count
- Completion rate percentage
- Average time to completion
- Median time to completion
Overall summary (AI-generated synthesis)
Overall themes (key themes identified across all conversations)
Overall opportunities (actionable insights)
Overall interesting quotes (notable participant quotes)
Per-question analysis:
- Question text and index
- Summary for each question
- Themes specific to each question
- Opportunities for each question
- Quotes relevant to each question
- Categories with labels, descriptions, counts, and percentages

File format: TXT only

Use cases:

Sharing analysis insights with stakeholders
Creating presentations and reports
Archiving AI-generated insights
Comparing analysis across different filters
Quick reference for key findings

File Formats

Excel (.xlsx)

Native spreadsheet format and the default choice. Recommended for most users.

Advantages:

Opens cleanly in Excel on a double-click — no import wizard, no separator settings, and no garbled accented or non-Latin characters, on any computer or language setting
Each export includes a Data Dictionary tab that explains every column (what it means and which version of the export introduced it), so the new presentation-order, answer-source, and stimulus columns are self-explanatory
Right-to-left languages (such as Arabic and Hebrew) display in the correct direction, cell by cell, even in a study that mixes languages

Use Excel for:

Everyday analysis and sharing
Opening reliably across different computers and regional settings
Studies in multiple languages, including right-to-left scripts

CSV (Comma-Separated Values)

Structured spreadsheet format. Choose this when a downstream tool (pandas, R, a BI tool, or a data pipeline) expects plain CSV.

File structure:

conversation_id,respondent_index,start_time,duration,age,gender
abc123,1,2025-01-15T14:30:00Z,00:20:00,25,Male
def456,2,2025-01-15T15:00:00Z,00:17:50,30,Female

Advantages:

Opens in Excel, Google Sheets, Numbers
Supports pivot tables and charts
Compatible with statistical software (SPSS, R, Python)
Automatic column alignment

Use CSV for:

Quantitative analysis
Statistical analysis
Charts and visualisations
Spreadsheet work

TXT (Plain Text)

Simple text format with tab or comma separation.

File structure:

conversation_id    respondent_index    start_time     duration
abc123            1                   2025-01-15     00:20:00
def456            2                   2025-01-15     00:17:50

Advantages:

Opens in any text editor
Lightweight
No formatting complications
Smaller file size

Use TXT for:

Manual review
Plain text archival
Import to qualitative analysis software
Minimal file size requirements

Export Process

Click Export (Transcripts table or single conversation) to open the export modal
Choose what to export (Answers in columns, Full conversation, or Summary & categories) — skipped for a single conversation
Select a format (Excel, CSV, or TXT) — Excel is selected by default
Click Export
The file is prepared and downloads to your browser's Downloads folder

Export Data Structure

Summary & Categories Columns

Standard columns:

conversation_id - Unique identifier
respondent_index - Participant number (1, 2, 3...)
summary - AI-generated summary of the conversation
start_time - ISO timestamp of when the participant sent their first response
duration - Time to complete in HH:MM:SS format (e.g., 00:19:55)
categories for Q1, categories for Q2, ... - Category labels assigned to each conversation per question

Metadata columns:

All metadata fields collected during the interview (e.g., participant_id, source, age, gender, language, custom screening answers)
Columns are dynamic and vary per project based on what metadata was captured

Full Conversation Columns

Each row represents one message. Conversation-level fields appear on the first message row only.

Columns:

conversation_id - Unique identifier
respondent_index - Participant number (1, 2, 3...)
start_time - ISO timestamp of first response
duration - Time to complete (HH:MM:SS format)
summary - AI-generated summary (first row only)
_id - Unique message identifier
role - "Interviewer" or "Participant"
created_at - Message timestamp
original_message - Message text content
translated_message_{language} - Translated content (one column per language if translations exist)
Metadata fields (first row only): all metadata captured during the interview (e.g., participant_id, source, age, gender)
question_id - Identifier of the question this message relates to (join key for SPSS/NVivo/Atlas.ti)
presented_order - The position at which this respondent saw the related question (reflects randomisation; blank for older data)
answer_source - voice (transcribed audio) or typed; blank for interviewer turns
media_count, media_files, media_types - Attachments on the message and their identifiers/types, in the order shown

Newer columns (question ID, presentation order, answer source, media) are added at the far right, so any existing column keeps its position and scripts keyed on the old layout keep working.

Example row structure:

conversation_id,respondent_index,start_time,duration,summary,_id,role,created_at,original_message,age
abc123,1,2025-01-15T14:30:00Z,00:12:30,AI summary text,msg001,Interviewer,2025-01-15T14:30:05Z,"Welcome!",25
abc123,,,,,msg002,Participant,2025-01-15T14:30:20Z,"Hello there",
abc123,,,,,msg003,Interviewer,2025-01-15T14:30:30Z,"What is your goal?",

Answers in Columns Export Columns

One row per respondent. The first row is the header; the second row repeats each question's full text as a label.

Columns, left to right:

conversation_id, respondent_index, start_time, duration, summary
One column per question (Q1, Q2, …) holding that respondent's answer, plus a translated column per question for each language when translations exist
categories for Q1, categories for Q2, … - Category labels per question
All metadata fields
Q1_presented_order, Q2_presented_order, … - The position (1st, 2nd…) at which this respondent saw each question; reflects question/group randomisation, blank for older data
Q{n}_stimulus_id, Q{n}_stimulus_label - For questions that show media, the identifier and label of the stimulus shown, in the order shown (only present for questions with media)

The presentation-order and stimulus columns are added at the far right so existing columns keep their positions.

Data Dictionary (Excel only)

Single Conversation TXT Format

Formatted as a readable conversation, with a header block, an optional summary and category breakdown, metadata, then the full transcript grouped by speaker:

================================================================================
TRANSCRIPT #1
================================================================================
Conversation ID: abc123
Respondent ID: participant-abc
Respondent Index: 1
Date: 2025-01-15 14:30:00
Duration: 20 minutes

SUMMARY:
    AI-generated summary of the conversation.

CATEGORIES:
  Q1: Convenience

METADATA:
  age: 25
  gender: Male
  location: New York
================================================================================

CONVERSATION TRANSCRIPT:

INTERVIEWER (2025-01-15 14:30:05):
    [P1]     Welcome! Let's begin. What is your main goal?

RESPONDENT (2025-01-15 14:31:00):
    [P1]     To save time

Working with Exports

Opening in Excel

Double-click CSV file, or
Open Excel → File → Open → Select CSV

The notes below cover rare cases where a file still needs a small adjustment, usually depending on your Excel version or system settings.

Dates display incorrectly:

Select date column → Format → Number → Date

Numbers not calculating:

Select column → Data → Text to Columns → Delimited → Comma

Special characters show as symbols:

This is rare. If it happens, re-open or re-import the file and set the
encoding to UTF-8.

Opening in Google Sheets

Open Google Sheets
File → Import
Upload tab → Select CSV file
Choose "Comma" as separator
Click Import

Analysis in Excel

Completion rate by demographic:

Create pivot table
Rows: demographic field (age, gender, etc.)
Columns: status
Values: count of conversations

Average duration by segment:

Create pivot table
Rows: segment variable
Values: average of duration

Keyword search in responses:

Use Find feature (Ctrl+F / Cmd+F)
Search the original_message column
Filter matching rows

Qualitative Analysis Software

NVivo:

Export as TXT
Import as dataset
Code transcripts
Run queries

Atlas.ti:

Export as TXT or CSV
Import as primary documents
Create codes
Analyse with quotation tools

Dedoose:

Export as CSV
Import as excerpts
Add descriptors from metadata
Code and analyse

MAXQDA:

Export as TXT
Import as documents
Use auto-coding features
Run text search

File Management

Naming Convention

Use descriptive filenames with dates:

Format: ProjectName-ExportType-YYYY-MM-DD.csv

Examples:

UserResearch-SummaryCategories-2025-01-15.xlsx
CheckoutStudy-FullConversation-2025-02-01.xlsx
MarketResearch-Age25to34-2025-01-20.csv

Include:

Project name/identifier
Export type (Answers in columns / Full conversation / Summary)
Date (YYYY-MM-DD format)
Filter criteria if applicable (Age25to34, Completed, etc.)

Folder Structure

Recommended organisation:

Project-Name/
  ├── Raw-Data/
  │   ├── 2025-01-15-SummaryCategories.xlsx
  │   ├── 2025-01-22-FullConversation.xlsx
  │   └── 2025-02-01-Final.csv
  ├── Analysis/
  │   └── Coded-Data.xlsx
  └── Reports/
      └── Final-Report.pdf

Export Log

Track exports in a spreadsheet:

Columns:

Date of export
Export type (Answers in columns / Full conversation / Summary)
Number of conversations
Filters applied
Purpose/milestone
File location

Example:

Date       | Type     | Count | Filters    | Purpose       | Location
2025-01-15 | Summary  | 23    | None       | Week 1 backup | /Raw-Data/
2025-01-22 | Full conversation | 47 | Age 25-34 | Segment data | /Analysis/
2025-02-01 | Full conversation | 75 | None     | Final archive | /Raw-Data/

Backup Schedule

Export data at:

Every 25 completed conversations
Weekly during active collection
Before deleting conversations
Before major project changes
Project completion

Filtering and Exports

How Filters Affect Exports

Active filters determine exported conversations:

No filter: exports all conversations in current view
With filter: exports only matching conversations
Filter criteria not included in export file name (document manually)

Segment Export Strategy

Export full dataset (no filters) as master backup
Apply filter to segment A
Export with descriptive filename
Remove filter
Apply filter to segment B
Export with descriptive filename
Repeat for each segment
Compare in analysis tool

Multiple filter combinations:

Example: Age 25-34 AND Gender: Female AND Location: California
Filename: Project-Women25-34-CA-2025-01-15.csv

Troubleshooting

Export button missing or disabled

Possible causes:

No conversations exist
Permission issues
Browser cache holding a stale UI state

Solutions:

Verify completed conversations exist
Confirm you have the required role (Owner, Admin, Editor, or workspace-level export permission)
Refresh browser
Log out and log back in

Download failed or corrupt file

Possible causes:

Network interruption
Browser blocking download
File too large

Solutions:

Retry download
Use different browser (Chrome or Firefox)
Check internet connection
Export smaller segments using filters
Clear browser cache

File won't open in Excel

CSV exports normally open in Excel on a double-click. In the rare case one
doesn't, try:

Right-click → Open with → Excel
Excel: Data → Get Data → From Text/CSV
Set encoding to UTF-8
Try Google Sheets

Scrambled data or weird characters

Exports display accented and non-Latin characters correctly out of the box, so this is rare. If you do see scrambled characters (for example after editing and re-saving the file in another tool):

Re-open or re-import the file with UTF-8 encoding
In Google Sheets: import with UTF-8 selected
Open in a text editor to verify the data
Re-export a fresh copy from Tellet

Missing conversations

Possible causes:

Active filter
Wrong view (Completed vs Incomplete)
Very recent conversations may take a moment to appear

Solutions:

Remove all filters
Check both Completed and Incomplete tabs
Refresh page
Verify conversations exist in Transcripts table

Data Privacy and Security

Data Sensitivity

Exports may contain:

Personal information
Demographic data
Opinions and feedback
Sensitive topics

Best practices:

Store in secure locations
Don't email unencrypted files with sensitive data
Use password-protected folders
Delete old exports when no longer needed
Follow organisational data policies

Sharing Exports

Before sharing:

Review included data
Confirm recipient needs access
Consider anonymisation or redaction
Use secure file sharing
Set link expiration dates

Anonymisation:

Remove participant identifiers
Generalise locations or details
Redact personal information in quotes
Use aggregate data when possible

Use Cases

Early Project (Exploration)

Export Summary & categories to check demographic distribution
Verify data quality
Create backup

Mid-Project (Analysis)

Export Full conversation for qualitative coding
Create pivot tables for segment analysis
Generate custom visualisations
Combine with other datasets

End of Project (Documentation)

Export complete dataset for archival
Share with external researchers
Meet data retention requirements
Final backup