Tackling the operational challenges of reducing churn and increasing ARPU
In order to tackle successfully the complex operational issues of reducing churn and increasing ARPU, telecom operators must adopt more granular and dynamic customer base analysis approaches allowing to address customer specific needs and concerns as they arise.
Innovative micro-segmentation approaches enable operators to detect customers at risk of churning and presenting a potential for upselling and cross-selling through focused pragmatic data approaches combined with machine learning that can continuously adapt to changing market and customer situations.
Designing and launching large volumes of personalized commercial and care actions across several channels can be implemented with the support of new productivity tools and more automation, such as personalized action recommendation engines.
The limits of traditional top-down segmentation
Customer segmentation was once viewed as the best practice for telecom operators to understand their customer base composition. Segmenting the customer base and analyzing performance per segment allowed operators to improve their marketing, sales and customer service efforts, as well as designing their organization and processes.
Initially, segmentation was elaborated on the basis of top-down market data and pre-defined analytical axes, which provided big picture views based on a limited number of broad categories. Over time, segmentation technics were refined with need-based and behavioral approaches, enabled by marketing tools such as customer surveys and focus groups. These top-down segmentation approaches however reached their limits and resulted in overly monolithic and static views of operators’ customer bases around a limited number of “manageable” predefined segments or personas – usually 7 to 9 segments, and often let to the creation of rigid processes and organizations that left operators ill-equipped to tackle evolutions in customer behaviors and market conditions. Moreover, these forms of rigid top-down segmentations are not well suited to efficiently address paramount operational issues such as churn, which require developing much more granular views.
More targeted approaches are now required
As operators have increasingly been focused on reducing churn and improving up/cross selling, they are developing more sophisticated segmentation approaches and design targeted actions directly addressing customers’ specific needs and concerns, which in turn requires more granular and dynamic analyses of their customer base beyond traditional top-down segmentations. Focusing on increasingly targeted customer actions thus requires new approaches and tools in order to elaborate more granular and flexible customer base segmentations and a greater level of automation to enable the execution of these targeted actions by different channels. Moreover, real time marketing has been rapidly developing, which also requires developing new approaches to leverage available customer data to nourish and automatize real time actions across both out-bound and in-bound channels.
Micro-segmentation can enable personalized actions
While traditional top-down macro-segmentation classifies customers according to a predefined set of hypotheses relating to the characteristics that are expected influence customer behaviors (e.g. demographics, usage information), an exploratory bottom-up approach to elaborate segmentation at a more micro level can be much more fruitful, allowing to reveal new variables that effectively determine customer behaviors and lead to develop approaches that addresses these customer behaviors. In order to elaborate the customer base granular view required to address customers with more targeted actions, operators must thus redefine their approach to customer segmentation, moving from top-down hypothesis-driven macro-segments to bottom-up data-driven micro-segments. Accordingly, operators must adopt new customer micro-segmentation approaches that are more focused on designing specific targeted actions, more flexible to address customer needs and concerns, and more evolutive to integrate market and customer situation changes.
The adoption of a well calibrated approach to micro-segmentation can allow telecom operators to move away from one-size-fits-all campaigns, which typically address large groups of customers relatively indiscriminately, and rigid customer journeys that do not necessarily match customer expectations. The objective of micro-segmentation is thus to enable improved retention and up/cross-selling through more targeted actions, which allows operators to address the concerns and needs of customers when they arise with specifically designed commercial and care personalized actions.
By way of illustration, developing a micro-segmentation customer base analysis approach to design and launch personalized actions can rapidly deliver churn rate reduction of 1 to 2.5 percentage point for an operator having already implemented strong retention practices.
The challenge of implementing micro-segmentation
Adopting micro-segmentation approaches entails important operational implications for telecom operators. The inherent problem of micro-segments is that they proliferate, which makes it difficult to operationalize micro-segmentation at scale to continuously feed channels with a large volume of targeted actions. Shifting from five to ten macro-segments to several hundreds of micro-segments raises real productivity, operational costs and IT system challenges, which can be solved by industrialized processes, new productivity tools and automation.
There are three major challenges for operators to adopt successfully a micro-segmentation approach targeting customers with specifically designed actions:
Leveraging data efficiently to discover micro-segments,
Adapting to rapidly evolving environments to maintain relevant micro-segments,
Automatizing targeted action design and launch to address micro-segments efficiently.
Leveraging data efficiently
First, having a readily structured access to customer data and the ability to efficiently exploit this data becomes crucial components of the operationalization of real targeted action approaches at a micro-segment level. This move to a more bottom-up customer base and customer behavior analyses necessarily depends on a more thorough exploitation of the massive volume of data dispersed within a large number of operator systems. Most telecom operators have been dedicating increasing time and resources to collecting, structure and develop data analytics and machine learning tools allowing them to exploit this massive volume of data they generate. Armed with these powerful data tools, the challenge for operators now lies in finding a balance between the desired contextualization/personalization for customers and the engendered complexity for their systems, processes and organization to deliver tangible results. This challenge can be addressed by starting to focus on a limited number of carefully selected OSS and BSS data items to develop a bottom-up approach, which can be progressively enriched with additional data items once initial results are obtained.
Adapting to rapidly evolving environments
Second, micro-segments must continuously evolve to remain relevant as new market trends emerge with new promotions and competitor initiatives, and customer situations and behaviors change over time (e.g. Usage changes, service incidents). Part of the difficulty is that predictive models and customer micro-segments design to address churn and up/cross-selling are generally developed by customer intelligence and marketing data analytics teams based upon customer data gathered for ad hoc projects. This process requires high human efforts from both marketing and data science teams to develop these predictive scoring models, including to update them on a regular basis, and this process often leads to the adoption of fixed micro-segments that tend to lose in relevancy as market conditions and customer situation evolve over time. Moreover, significant additional analytical work is typically required to generate the analysis allowing marketing, product, customer experience and sales teams to turn analytical results (typically detailed scoring) into concrete and impactful commercial and care actions.
These challenges can be addressed by resorting to new innovative autonomous clustering approaches, which can be regularly trained to learn out of actualized market and customer data. Clustering models can enable operators to automatically detect micro-segments of customers that are at-risk of churning or that have a potential for up/cross-selling, without starting from a top-down or pre-defined approach. This automatic micro-segment discovery can be retrained automatically on a regular basis (e.g. every month) in order to always be up to date: micro-segments can thus evolve over time as customer situations and market conditions constantly change.
Automatizing personalized action design and launch
Third, some level of automation is required in order to enable marketing, customer experience and sales teams to efficiently design and launch personalized actions over a large volume of customer micro-segments. Indeed, the shift from macro-segments and large volume campaigns to micro-segments and targeted actions entails a considerable surge in the types of actions to be designed and the volume of actions to be launched. This is a productivity challenge that automated recommendation engines can help to match the micro-segment with the best type of actions.
Leveraging cluster discovery models, action recommendation engines can propose both actions based on pre-defined marketing rules (top-down approach) and actions based on past performance with similar customers (bottom-up approach).
Accordingly, for each micro-segment discovered by the clustering model, automatic action recommendations can be based on:
the matching of actions with customers belonging to a given micro-segment based on operators’ marketing rules, and
the similarities between customers belonging to a given micro-segment and, for each action, the profiles of the customers having widely accepted the same type of actions, or improved their survival rate and subscription ARPU subsequent to the action.
Customer base micro-segmentation enables operators to address the concerns and needs of customers when they arise with specifically designed commercial and care personalized actions allowing to reduce churn and increase ARPUs. Adopting customer base micro-segmentation to deliver personalized actions requires that operators adopt innovative and pragmatic approaches: a focused data approach that can be readily implemented, machine learning that continuously adapts to evolving market and customer situations, and automatized action design and launch powered by new tools and processes.
How Lifetime Analytics enables operators to manage their customer base
Lifetime Analytics has recently launched an application providing telecom operators with a productized solution to design, manage and optimize micro-campaigns targeting customer micro-segments with personalized commercial and care actions across different channels. This new application is designed to improve retention and up/cross-selling through more targeted actions that are recommended by the application, allowing to address the concerns and needs of customer micro-segments when they arise with specifically designed commercial and care actions. Moreover, these automatic micro-segment discovery and action recommendations are retrained every month in order to always be up to date: micro-segments accordingly evolve over time as customer situations and market conditions constantly change.
Using the Lifetime Analytics application, marketing, product, sales and customer experience teams can
rapidly identify and act on updated root causes of churn/down-selling and up/cross-selling with powerful analytical insights,
leverage action recommendations for each customer profile,
launch and track the performance of several actions to reduce churn and increase ARPU.
Moreover, the Lifetime Analytics application complements the scheduled CRM campaigns activities with an over-the-top solution to efficiently orchestrate micro-campaigns/actions through a single end-to-end application, with significant productivity gains for marketing, product, sales and customer experience teams.
By Frédéric BEAUVAIS and Julien CABOT
The authors are co-founders of Lifetime Analytics S.A.S.