The mobile phone operator Amena has begun an important successful project aimed at retaining and loyalizing their residential clients.

Technical Data

Company: Amena.

Objectives: define client profiles and predict
their tendency to cancellation (scoring).

Solution: create a client retention model in
the Residential Business Unit.

Result: adjust marketing and loyalization
campaigns to each client profile.

Knowing client’s behavior and preventing the risk of abandonment has its benefits. Not only can it translate in terms of economic performance and increased income, but it can also improve spending and optimization of available resources. In fact, for Amena, it was an opportunity to face this problem with the goal of launching an appropriate portfolio policy, adjusting their marketing ingredients according to the type of consumer the campaigns were aimed at.

As a preliminary step, the company needed in-depth knowledge of the profile for residential clients that had previously cancelled their service, so that they could establish a set of rules applicable to each one of the identified groups, as well as the tendency toward cancellation for the classes involved. The analysis, which was done throughout 2004, focused on the company’s residential clients. When trying to determine the probability of abandonment by segments, we started by defining the term “leave”. Then three categories were distinguished:

• Voluntary leaving (originating in the client and caused by dissatisfaction with service, the existence of better offers from the competition, or other external motives).
• Involuntary leave (missed payments or fraud).
• Porting leave (contracts from clients that had requested a change to another operator).

Next, the necessary information was gathered to establish a data mining model, analyzing the residential clients that had cancelled service and others who had not, and determining, in both cases, the characteristics of each, in such a way that it was possible to distinguish the different existing profiles to be able to compare them and establish the parameters that end up explaining the consumer’s future behavior. The data used in this part of the project responded to different typologies:

• Contract data: seniority, type and prices.
• Telephone use profile or pattern of use: total use, type of calls, days on which service was used, time of day calls are made, etc.
• Socio-demographic data: age, sex, client’s area of activity, etc.
• Link to the company: number of services hired and situation of each one.
• Other data: number and type of contacts and incidences gathered in the Customer Service Center.

With all of that, a series of rules was established with which to identify general profiles, singling out each one using data such as their size or their rate of cancellation probability. Then, a scoring system was developed, capable of assigning an abandonment probability indicator to each client group automatically. In this way, it is possible to assign a cancellation probability number to each contract or client, depending on the case. This allows us to establish criteria to organize clients according to this indicator, and afterwards we can take more precise loyalization actions. Finally, churn studies were done, focusing on the residential segment, which was the object of the project.

With all of this methodology, we were able to establish an order of action by profile that allows us to carry out personalized loyalization campaigns (Marketing 1-to-1) that are more selective and efficient, and are also adjusted to the needs of each consumer group.

The work undertaken has allowed the company to make intelligent and rational use of their resources, developing different loyalization strategies for each user profile, which resulted in savings and a notable improvement in sales efficiency. In this way, in-depth knowledge of abandonment motives has served to improve client retention, adjusting the offer to their specific needs.