Pricing

Rating

Developer

Actor stats

Categories

Share

🔍 Churn Scout - Market Intelligence Agent

Autonomous AI-powered market intelligence agent that reveals why customers are leaving your competitors

👁 Apify
👁 Python
👁 ML Powered

🎯 What is Churn Scout?

Churn Scout is an autonomous market intelligence agent designed for SaaS founders, product managers, and marketing teams. It automatically crawls multiple public developer and social channels (Hacker News, GitHub Issues, DEV.to, StackOverflow) to identify customer frustrations and churn signals about a target competitor.

Unlike basic text-dumps, Churn Scout utilizes an internal Machine Learning Engine (Sentiment Analysis + TF-IDF Vectorization + K-Means Clustering) to group complaints into specific Pain Point Categories (e.g., "Pricing Issues", "Performance Problems"). It compiles these signals into a premium, self-contained, interactive HTML dashboard containing visualized metrics, client-side filters, and strategic competitor positioning plans.

✨ Features

🎛️ Advanced Target Crawling

• Multi-Source Support: Scrape Hacker News, GitHub Issues, DEV.to, and StackOverflow concurrently.
• Custom Keywords: Append custom search phrases to look for specific complaints (e.g. "slow", "bug", "pricing").
• Resilient Architecture: Automatic retries and exponential backoff handling rate limits gracefully.

🧠 Intelligent Processing

• Sentiment Filtering: Automatically filters out noise to target high-frustration indicators (polarity below a user-defined threshold).
• Automated Categorization: Uses statistical NLP to cluster evidence into clear pain areas.
• AI Strategic Insights: Add your API key (Gemini, OpenAI, or OpenRouter) to generate executive summaries, tactical roadmaps, and competitive pitches.

📊 Premium Visual Dashboard

• Modern Aesthetics: Built with a sleek dark-mode glassmorphic theme inspired by Stripe and Vercel dashboards.
• Interactive Visualizations: Embedded Chart.js charts showing sentiment split and category frequency.
• Client-Side Search & Filters: Live search over signal text and tabs to filter by platform or severity.
• Data Portability: Clean export buttons to download filtered datasets directly to CSV or JSON.

🚀 How It Works

graph TD
 Input[Competitor Name & Settings] --> Scraper[Resilient Multi-Source API Scraper]
 Scraper --> Sources[HN, GitHub, DEV.to, StackOverflow]
 Sources --> Sentiment[TextBlob Sentiment Score]
 Sentiment --> Filter[Sentiment & Noise Filter]
 Filter --> ML[TF-IDF Vectorization + Clustering]
 ML --> AI[AI Strategic Positioning Generator]
 AI --> Dashboard[Interactive HTML Dashboard]
 AI --> Dataset[Apify Dataset Output]

📥 Input Configuration

Example Input

{
"competitorName":"Slack",
"maxPosts":150,
"sources":["Hacker News","GitHub Issues"],
"minSentiment":-0.1,
"customKeywords":"expensive, mobile app",
"apiKey":"YOUR_GEMINI_OR_OPENAI_API_KEY",
"proxyConfiguration":{
"useApifyProxy":true
}
}

📤 Output

1. Interactive Dashboard (HTML)

Stored in the default Key-Value Store under the key OUTPUT. Contains:

• KPI Metrics: Churn signal counts, Average Sentiment, Disruption vulnerability level.
• Visual Analytics: Interactive bar charts (pain point frequencies) and doughnut charts (sentiment distribution).
• Tactical Strategy: Positioning advice, immediate marketing options, and product differentiation opportunities.
• Raw Evidence Table: Filterable, searchable, and sortable table of all scraped complaints.

2. Dataset Records (JSON / CSV)

Includes structured metadata for every signal:

• topic: Clustered pain point category.
• text: Raw complaint text.
• polarity: Raw sentiment polarity (-1.0 to 1.0).
• source: The source site.
• date: Scrape item publishing date.
• engagement: Points, comments, or reaction count.
• url: Link directly to the community thread.

🛠️ Technology Stack

• NLP: TextBlob Polarity Scoring
• Machine Learning: Scikit-Learn (TF-IDF Vectorizer, K-Means Clustering)
• Engine: Python 3.11 + Apify SDK
• Visualization: Jinja2 Templates, Chart.js, Lucide Icons

🔒 Compliance & Safe Harbor Guidelines

This Actor is built to operate under strict compliance criteria:

• Public Data Only: Accesses public APIs and indexing paths without logging in.
• Rate Limit Adherence: Employs query limits and backoff pauses to protect target servers.
• Privacy-First: Focuses purely on product critique; strips PII and user-specific details.
• Transformative Utility: Creates high-level aggregated strategic business intelligence.