The primary stakeholder is EZTABLE, an ecommerce based platform for reservations for different restaurants in Taiwan. Currently, there is a mismatch between the number of bookings and the amount of revenue generated by the bookings. The problem is the current booking system which enable the user to book reservations even without prior payment. Hence the actual revenue falls short of the potential revenue due to the existing system. The business goal of our project is to let our client realize the potential revenue and the need for changing the strategies. At the same time, the mode of payment is changing its trend from CASH to ATM. It gives a clear indication that the consumer habits are changing rapidly. Undoubtedly, this is the right time for our client to take some strategic decisions which can help it to achieve the potential revenue.
The data consists of the purchases information for EZTABLE. (Mention the time and number of records). We handle the missing value and outlier value. We aggregate data into monthly one and partition it. All we do before forecasting is to make our forecasting more accurate. Since the revenue has both trend and seasonality, we have used the methods which can take care of the both.
We choose the multiple linear regression to do the forecast because the data has polynomial trend and seasonality. The input values are T, T square, seasonal index and the output value is payments. Our client can use the forecasting model to predict the actual revenue and potential revenue.
We suggest EZTABLE to re-design the payment systems which is more convenient for customer to pay. To make sure the reservation is successful, our client can do targeted marketing such as if they use the booking then they might receive some offer, etc. In this way, EZTABLE can catch up potential revenue by increasing its actual revenue.