Purpose of clusters in Lifetime Analytics

A key concept of the Lifetime Analytics for a better loyalty, cross-selling or up-selling in Telecoms is to work per Product family and per cluster, instead of segments or using a scoring.

The purpose of the cluster is to work with sets of subscription having a common "profile" regarding the churn, cross-selling and up-selling. The "profile" can be same usage, same customer experience, same allowance, same socio/local properties, same billing range, etc and a mix of those.

In Lifetime Analytics, you can work with :

  • Churn cluster : set of subscribers presenting the same common profile to churn in the next months (see Horizon), uplifting the LTM churn rate over the product family and to track the impact of actions on the cohorts through the survival rate.
  • Up-sell cluster : set of subscribers presenting the same common profile to accept up-selling offers - migrating to another Marketing product  into the same Product Family with an increment of median MRC of at least 2€ (see configuration to adjust) - in the next months (see Horizon), uplifting the MRC ARPU LTM over the product family and to track the impact of actions on the cohorts through the MRC ARPU growth rate.
  • Cross-sell cluster : set of subscribers presenting the same common profile to accept cross-selling offers - adding another subscriptions - in the next months (see Horizon), uplifting the Customer MRC ARPU LTM over the product family and to track the impact of actions on the cohorts through the Customer MRC ARPU growth rate.

See the Main Menu to access to the different clusters per type.

Properties of a cluster

A cluster apply to a single Product Family with an Horizon period in months (1 to 12 months). The Horizon is the term of the targeted event (churn, cross-sell or up-sell). Eg Short term churn clusters, with Horizon equal to 1 month, are the the sets of subscribers presenting profiles for the churn in the next months. The Horizon defines into the Cluster details charts when to look about the subscription data.

A cluster is defined by a rule in natural language to help the understanding by the Marketing team, as a set of mutual conditions (logical AND) based upon the Data Reference to select the subscription belonging to the cluster. Note : a subscription can belong to different clusters. The accumulated revenue or volume are de-duplicated in the application to avoid double counting.

The main benefit of the Lifetime Analytics application is its ability to discover automatically the best clusters from the Subscription data for churn, cross-sell and up-sell. Additionally, you can create or customize your own custom clusters (See Create Custom Clusters).

Designing an Action Plan is performed on subset of the cluster, using some filters (See Design Action Plan)

Churn cluster

Churn clusters are sets of active subscribers presenting the same profile as past churners, having churn at N+ horizon months. These subscriptions are defined as subscriptions at risk (of churn) when there are free of commitment or having at least one missed payment in the last 3 months.

See Deep Dive Into Churn Cluster Details for more information

 

Up-sell cluster

 

Up-sell clusters are sets of active subscribers presenting the same profile as past up-sellers, having up-sell at N+ horizon months, from the product family. These subscriptions are defined as subscriptions to target to up-sell to marketing product producing a greater average monthly revenue charge (MRC).

See Deep Dive Into Up-sell Cluster Details for more information

In the application, Family up-sell clusters are a particular type of clusters for migrations between two product families.