Opera Browser -- Engaged Users Survey Analysis

Open-ended text analysis of Opera Browser's most engaged desktop users, covering four critical survey questions about browser switching, recommendation drivers, and churn risk. Built for the Opera Growth & Product team to understand what keeps power users loyal, what threatens retention, and what drives word-of-mouth advocacy.

---

What You'll Find Here

For a human reader: This repository contains raw open-ended survey responses from Opera's engaged desktop user base (the "Engaged Users Survey", June 2025) alongside Python scripts that generate formatted Word document summaries of each question's analysis. The four questions cover the full retention lifecycle: why users might switch away (Q14, 8,592 responses), why they recommend Opera (Q17, 1,463 responses), what would make them leave entirely (Q32, 1,922 responses), and what specifically drove their recommendations (Q35, 223 responses). Each analysis uses semantic clustering to surface the top themes, backed by real user quotes in multiple languages. The pre-built DOCX reports are included so you can review findings without running code.

For an LLM or automated system: The four CSV files contain structured open-ended survey microdata with columns for Respondent ID, Response Date, and free-text Responses/Tags. The text spans 8 languages (English 57%, Portuguese 26%, German 9%, French 5%, plus Spanish, Polish, Italian, and mixed). Each Python script contains hardcoded analysis results (theme rankings, percentages, representative quotes) encoded as data structures and generates a formatted DOCX using python-docx. The data can be used for further NLP analysis, sentiment classification, topic modeling, or as ground truth for training text classifiers on browser user feedback. The markdown summary (Opera_Browser_Switch_Analysis_Summary.md) provides a machine-readable version of the Q14 analysis.

---

Dates and Provenance

Detail	Value
Survey period	June 2025
Analysis scripts created	February 20, 2025
DOCX reports generated	June 24, 2025 (Q14, Q17) / February 20-25, 2025 (Q32, Q35)
Survey source	Opera internal "Engaged Users Survey" -- desktop power users
Total responses across all questions	12,200 open-ended text responses
Languages represented	English, Portuguese, German, French, Spanish, Polish, Italian
Analysis method	Semantic clustering and thematic categorization via batch processing

---

Survey Questions Analyzed

Question	CSV File	Responses	Core Question
Q14	Q14_Text.csv	8,592	"What could be the reason for you to switch to another browser?"
Q17	Q17_Text.csv	1,463	"What was the reason for your recommendation?" (why users recommend Opera)
Q32	Q32_Text.csv	1,922	"What could be the reason for you to leave Opera?"
Q35	Q35_Text.csv	223	"What was the reason for your recommendation?" (detailed follow-up)

---

Key Findings at a Glance

Q14 -- Browser Switch Reasons (8,592 responses)

Rank	Theme	%	Responses
1	Performance issues (RAM, CPU, crashes)	26.7%	2,293
2	Feature dissatisfaction (forced AI, bloat)	18.0%	1,547
3	Website compatibility (banking, streaming)	14.0%	1,202
4	Loyalty / no reason to switch	12.0%	1,030
5	Security & privacy concerns	8.0%	687

Q17 -- Recommendation Reasons (1,463 responses)

Rank	Theme	%	Responses
1	Sidebar features (WhatsApp, Aria, messengers)	29.0%	424
2	Tab management (islands, workspaces)	28.0%	410
3	Video/multimedia (picture-in-picture)	15.0%	219
4	Chrome alternative (anti-big-tech)	12.0%	176
5	General usability & interface	8.0%	117

Q32 -- Churn / Leave Reasons (1,922 responses)

Rank	Theme	%	Responses
1	No reason to leave / loyal users	29.0%	557
2	Performance & stability issues	22.0%	423
3	Feature-related concerns (VPN removal, forced AI)	15.0%	288
4	Privacy & security concerns	13.0%	250
5	Website compatibility issues	8.0%	154

Q35 -- Recommendation Drivers (223 responses)

Rank	Theme	%	Responses
1	Sidebar features & integrations	29.1%	65
2	Chrome alternative positioning	17.9%	40
3	Workspaces & tab management	15.7%	35
4	Performance & stability	13.9%	31
5	User-friendly design	12.1%	27

---

File Summaries

Data Files (CSVs)

Each CSV has a header comment line (the question text), then columns: Respondent ID, Response Date, Responses (or Other (please specify)), Tags.

File	Size	Rows	Survey Question	Description
Q14_Text.csv	988 KB	8,592	"What could be the reason for you to switch to another browser?"	Largest dataset. Free-text responses about potential browser-switching triggers. Multi-language (8 languages).
Q17_Text.csv	120 KB	1,463	"What was the reason for your recommendation?"	Responses from users who had recommended Opera to others. Captures advocacy drivers.
Q32_Text.csv	160 KB	1,922	"What could be the reason for you to leave Opera?"	Churn-focused question. Similar to Q14 but framed around leaving Opera specifically (not switching to a named competitor).
Q35_Text.csv	19 KB	223	"What was the reason for your recommendation?"	Smaller follow-up dataset on recommendation drivers. More detailed responses from a subset of advocates.

Analysis Scripts (Python)

Each script uses python-docx to generate a formatted Word document containing the full analysis: executive summary, methodology, language distribution table, ranked themes with percentages and user quotes, strategic insights, and recommendations.

Script	Output	Question	Responses Analyzed	Top Finding
create_docx_summary.py	Opera_Browser_Switch_Analysis_Summary.docx	Q14	8,592	Performance issues are the #1 switch reason (26.7%)
create_q17_docx_summary.py	Opera_Q17_Recommendation_Analysis_Summary.docx	Q17	1,463	Sidebar features are the #1 recommendation driver (29.0%)
create_q32_docx_summary.py	Opera_Q32_Churn_Analysis_Summary.docx	Q32	1,922	29% of users have no reason to leave (strongest loyalty signal)
create_q35_docx_summary.py	Opera_Q35_Recommendation_Drivers_Analysis_Summary.docx	Q35	223	Sidebar integrations again lead at 29.1%

Generated Reports (DOCX)

Pre-built Word documents with formatted tables, headings, bullet lists, and styled quotes. Can be opened directly without running any code.

File	Size	Contents
Opera_Browser_Switch_Analysis_Summary.docx	40 KB	Q14 analysis: top 10 switch reasons, language breakdown, business insights, strategic recommendations
Opera_Q17_Recommendation_Analysis_Summary.docx	40 KB	Q17 analysis: top 10 recommendation reasons, unique selling propositions, critical success factors
Opera_Q32_Churn_Analysis_Summary.docx	40 KB	Q32 analysis: top 9 churn factors, retention strategies, highest churn risks vs. retention factors
Opera_Q35_Recommendation_Drivers_Analysis_Summary.docx	31 KB	Q35 analysis: top 10 recommendation drivers, cross-dataset consistency notes, statistical summary

Markdown Summary

File	Contents
Opera_Browser_Switch_Analysis_Summary.md	Machine-readable markdown version of the Q14 analysis. Contains the same content as the DOCX (executive summary, all 10 themes with quotes, business insights, strategic recommendations). Useful for direct consumption by LLMs or for rendering in GitHub.

---

How to Run

Prerequisites

Python 3.8+
python-docx

Install dependencies:

pip install python-docx

Generating a Report

All scripts expect to be run from the repository root (same directory as the CSV files):

cd engaged-users
python create_docx_summary.py

Each script generates its corresponding .docx file in the current directory, overwriting any existing file with the same name.

Generating All Reports

for script in create_*.py; do
    echo "Running $script..."
    python "$script"
done

---

Repository Structure

engaged-users/
├── README.md
├── .gitignore
│
├── Q14_Text.csv                                           # Switch reasons (8,592 responses)
├── Q17_Text.csv                                           # Recommendation reasons (1,463 responses)
├── Q32_Text.csv                                           # Churn / leave reasons (1,922 responses)
├── Q35_Text.csv                                           # Recommendation drivers (223 responses)
│
├── create_docx_summary.py                                 # Q14 analysis -> DOCX
├── create_q17_docx_summary.py                             # Q17 analysis -> DOCX
├── create_q32_docx_summary.py                             # Q32 analysis -> DOCX
├── create_q35_docx_summary.py                             # Q35 analysis -> DOCX
│
├── Opera_Browser_Switch_Analysis_Summary.docx             # Q14 report (Word)
├── Opera_Q17_Recommendation_Analysis_Summary.docx         # Q17 report (Word)
├── Opera_Q32_Churn_Analysis_Summary.docx                  # Q32 report (Word)
├── Opera_Q35_Recommendation_Drivers_Analysis_Summary.docx # Q35 report (Word)
│
└── Opera_Browser_Switch_Analysis_Summary.md               # Q14 report (Markdown)

---

License

Internal Opera Software analysis. Not intended for public redistribution of the underlying survey data.