I was awarded FIRST prize for this presentation and analysis

Complete data analyis of 3 hotels of MORGENS

This analysis was prepared for the DATA36 education platform’s December 2024 challenge. My work was selected among the top 3 highest-rated analyses, which allowed me to present it in front of a jury of senior data scientists and an audience on December 11, 2024.

Dataset

The analysis was conducted using 11 CSV files per hotel, covering 3 hotels in total. The dataset contains fresh, real-life data on hotel occupancy, realized bookings, customer searches, marketing campaigns, and the amount spent on each marketing channel.

Main questions

• Analysis of marketing elements (sources, mediums, campaigns) and conversion performance.
• Comprehensive analysis of search, booking, and occupancy data.
• Based on search, booking, and occupancy data, is it possible to develop a warning system to monitor marketing costs and maximize occupancy while maintaining optimal pricing?

Key Findings

• Bookings and searches are mainly concentrated around holiday periods (school autumn break, Christmas/New Year) and weekends.
• Booking window: Guests make reservations 3–4 months in advance. Two peaks exist: 2–3 days before booking and 96–97 days in advance.
• Conversion: 2–9% of searches for specific dates result in actual bookings.
• Guest types: Families with children and Hungarian guests prefer longer advance planning.
• Average stay: 2.55 nights; families with children search for longer stays but actually stay shorter periods.
• Room types: Three are the most popular (EE, FF, GG), but families with children prefer the other three types (AA, DD, CC).
• Marketing channels:
  • Google: most cost-effective, strong correlation with paid visitors (79%)
  • Meta: least cost-effective, correlation 92%
  • Microsoft: weak correlation with paid visitors (11%)
• All three channels have moderate correlation with bookings (38–46%)
• Campaigns: 8 major campaigns drove the majority of paid visitors.
• User behavior on website: visitor → search → booking.

Optimization Recommendations

• Advertising cost per booking ≤ 5% or 10,000 HUF per booking; intervene if exceeded for 3 consecutive days.
• Continuous measurement: track conversions weekly/monthly; aim to convert more website visitors into searchers and searches into bookings.
• Focus on the best-performing platform (for Hotel 1: Google).
• Remarketing: target users who started but did not complete bookings.
• Campaign planning:
  • For Hungarian audience: start campaigns 3–4 months in advance
  • For foreign audience: shorter lead time is sufficient
  • For families with children: consider campaigns supporting longer stays