Information Technology Reference
In-Depth Information
customers can only be frequent flyers because the rest of the people who are not
frequent flyers are just passengers, treated simply as numbers. What the concept
of identifying customers refers to in this section is the identification of a valuable
customer from the whole available set of passengers irrespective of whether they are
frequent flyers or not.
Identifying customers first entails that each passenger that has ever made a book-
ing with the airline is identified with a unique number. The definition and measure-
ment of the value of the customer can be defined by the airline business analyst,
however now the analyst is not restricted to members of the frequent flyer program
alone. This value of a customer can be measured in different ways, amongst them
travel frequency, revenue generated and the number of social ties the passenger has.
Identifying potential valuable customers from the whole spectrum of passengers can
prove beneficial for the airline to identify previously unknown potential valuable
customers and build lasting profitable relationships with them.
Apart from targeting valuable customers to join the frequent flyer program, this
information can be easily used to improve current customer support. A common
example is when a passenger forgets to provide his frequent flyer number at the
time of booking a flight. The system can recognise that the passenger is already
a frequent flyer and interact with the frequent flyer system to notify it of the sale
without the intervention of the customer, thus providing a better level of service.
5.2
Features of the Data
In order to study the extent of missing data which is not compulsory, a sample of
over 200,000 records was analysed. Since we are concerned with uniquely iden-
tifying entities, therefore passengers, the fields of interest were mainly those that
contain the passenger contact details. This data sample is specific to one airline,
however airlines that operate on a similar business model have a similar distribu-
tion. Table 4 shows the results of this analysis.
Ta b l e 4
Missing records in each field
Field Percentage Missing
Address 39.9%
Zip Code 39.9%
Frequent Flyer No 65.4%
Phones
75.8%
Email Address
78.6%
Title
93.7%
Group Name
95.3%
The only data element that uniquely identifies a passenger is the frequent flyer
number, however only 35% of the passengers in dataset have a frequent flyer num-
ber. Passengers usually find value in enrolling in frequent flyer programs because
