Personalized News & Podcast Platform
Mobile-first news platform with AI-curated content feeds, integrated podcast experience, and subscription management for independent journalism
Overview
Newslaundry is an independent, subscriber-funded news organization challenging mainstream media narratives. We built their mobile platform with personalized content recommendations that balance user preferences with editorial integrity, an integrated podcast player that became central to their content strategy, and subscription features that drove sustainable revenue growth.
The Problem
Independent journalism faces a discovery problem. Quality content exists but gets buried by algorithmic feeds optimizing for engagement over substance. Newslaundry had loyal web readers but struggled to reach mobile-first audiences and convert them to paying subscribers. Their podcasts were scattered across third-party platforms, missing the opportunity to drive engagement and subscriptions.
Key Challenges
Algorithmic vs Editorial Balance
Pure algorithmic recommendations optimize for clicks, not journalism quality. But zero personalization means readers miss content they'd value. The challenge was building AI that respects editorial judgment while learning user preferences.
Podcast Fragmentation
Newslaundry produces 10+ podcast shows but listeners consumed them on Spotify, Apple Podcasts, etc.—platforms that don't share listener data or drive subscriptions. No unified listening experience or cross-show discovery.
Subscription Conversion
Free readers needed compelling reasons to subscribe. The value proposition wasn't clear in the web experience. No mobile-native subscription flow with easy payment methods (UPI, cards, wallets).
Offline & Bandwidth
Many readers in India have inconsistent connectivity. Long articles and podcasts need offline availability. App size and data consumption matter for users with limited data plans.
How We Built It
Content Platform & Feed Architecture
Built headless CMS integration pulling articles and podcasts into unified content API. Designed feed architecture supporting multiple ranking algorithms (recent, trending, personalized, editorial picks). Implemented content tagging system for recommendation engine.
Podcast Player & Experience
Developed custom podcast player with background playback, sleep timer, and variable speed. Built cross-show discovery and personalized podcast recommendations. Implemented download manager for offline listening with smart storage management.
Personalization Engine
Created hybrid recommendation system combining collaborative filtering (similar users' preferences) with content-based filtering (topic similarity). Added editorial boosting mechanism allowing editors to surface important stories regardless of predicted engagement.
Subscription & Engagement
Built native subscription flows with Stripe and Indian payment methods. Implemented paywall logic with metering (X free articles/month). Created engagement features including saves, reading history, and push notifications for followed topics.
Solution Highlights
Hybrid Personalization
AI recommendations that learn user preferences while respecting editorial priorities. Editors can boost important stories that might not perform well algorithmically. Transparency into why content is recommended.
Unified Podcast Experience
All Newslaundry shows in one place with seamless playback, offline downloads, and cross-show discovery. Background listening, sleep timer, and variable speed for commute-friendly consumption.
Subscriber-First Features
Exclusive content clearly marked. Subscription flows optimized for Indian payment preferences. Subscriber-only features like ad-free experience and early access to shows.
Offline-First Design
Articles and podcasts downloadable for offline consumption. Smart storage management suggests what to remove when space is low. Low-data mode for bandwidth-conscious users.
Technical Deep Dive
The recommendation engine uses a two-stage architecture: candidate generation (fast, approximate retrieval of relevant articles using embedding similarity) followed by ranking (precise scoring using user features, content features, and contextual features). Editorial boosting is implemented as a score multiplier applied post-ranking, allowing editors to surface content without completely overriding personalization. The podcast player uses ExoPlayer on Android and AVFoundation on iOS with a shared React Native UI layer. Download management uses a priority queue with network-aware scheduling.
AI Capabilities
Content Personalization
Learning reading preferences without creating filter bubbles
Topic Extraction
Automatic tagging of articles for recommendation and discovery
Engagement Prediction
Forecasting which content will resonate with which user segments
Churn Prediction
Identifying subscribers at risk of cancellation for retention outreach
Technology Stack
Mobile
Backend
ML/Personalization
Audio
Payments
Results & Outcomes
4x
Subscriber growth
Year-over-year increase in paid subscriptions
45 min
Daily engagement
Average time spent per active user
500K/mo
Podcast listens
Monthly plays across all shows
4.8★
App rating
Average across iOS and Android stores
60%
In-app subscriptions
Share of new subscriptions originating from app
-40%
Churn rate
Subscriber retention improvement with engagement features
Services Used for the Client Product
Looking to solve similar challenges in your industry? Our team combines deep technical expertise with industry knowledge to deliver AI-powered solutions that drive measurable results.
Start Your Project
Let's discuss how we can help transform your operations with AI-powered solutions.
Continue exploring
See more case studies