Marketing Analytics for a Travel Company
One of the top U.S. travel companies, uses Machine Learning based predictive model to drive its entire marketing strategy. Model combines LifeTime Value prediction with web browsing, telephone and email activity of the customer to identify right products and messaging for the customer.
Business Impact
Improved marketing effectiveness, reduced annual marketing cost by 20% ($20M to $16M).
Grew annual passengers served by 20% (95k to 114k).
Increased profitability by reducing discounts.
Increased customer loyalty by more coherent marketing strategy
Methodology
A product driven marketing strategy is used, by identifying a set of customers to market to each week, for a set of products. A Machine Learning driven strategy is used to change the focus from product, to customers. Steps followed to make this model are:
First, a logistic regression model was made to score each interaction. Started with scoring each customer across various touch points like website browsing data, call center data, physical survey data and events data.
Second model aggregates each interaction at a customer-product level and ranks products for each customer. This is used to identify the right products to market to each customer.
Third model aggregates scores from previous model and predicts likelihood of booking a trip within next 30 days for each customer.
This model helped understand where a customer is in their booking cycle, i.e., whether they are ready to make a purchase, finalizing a trip from shortlisted 2-3 trips or at an early stage of exploring trips.
