Basel Alebdi & Nouf Alroqi- Data-Driven F&B Delivery: Jahez as a Leading Example| PyData Global 2023

Discover how Jahez, Saudi Arabia's leading food delivery platform, leverages Python, NLP & ML to optimize search, predict cuisines & analyze customer behavior for better delivery experiences.

Key takeaways
  • Jahez is a major food delivery platform in Saudi Arabia expanding internationally, with 600,000+ daily active customers and 25,000+ restaurant branches

  • The platform’s Data & AI department, established 1.5 years ago, has created products that generate 15% of total orders and 20% of revenue from newly joined restaurants

  • Search functionality accounts for 55% of Jahez’s orders, making it a critical feature for the platform

  • The company uses NLP and machine learning to:

    • Predict restaurant cuisines automatically
    • Identify missed business opportunities through search analysis
    • Correct spelling mistakes and match similar queries
    • Handle multi-language (Arabic-English) search capabilities
  • Customer behavior analysis shows that users leveraging data-driven features experience a 26% decrease in average order cycle time

  • Platform generates extensive operational data, with drivers covering 450 million kilometers (equivalent to 1000 moon trips)

  • Restaurant insights platform provides merchants with:

    • Customer base analytics
    • Operational performance metrics
    • Menu optimization suggestions
    • Basket analysis
    • Item preparation timing data
  • Search query analysis includes:

    • Substring matching
    • Levenshtein distance calculations
    • Session-based grouping
    • Multi-language name matching
  • System supports multiple cuisine classifications per restaurant to handle establishments serving different types of food

  • Data-driven approach has led to hundreds of new restaurant partnerships within two months of implementation