“Where’d You Go?” or a Quest for Churn Reduction

_{Listen to this post}

“If you don’t like your Internet provider, just switch it,”

— some guy from Reddit.

With the UK telecoms market being saturated and EE, O2, Vodafone and Three owning 82% of the mobile market combined, it gets progressively harder for the new players to enter the scene, while the opportunities for the ones already in the game are diminishing day to day. In the US, the situation is no better: major providers such as AT&T, Verizon and T-Mobile take up 98.5% of all wireless subscriptions according to Statista. Virtually any person has a connectivity service provider, and most residential areas have all of the major ones available which make it easy for the users to switch.

Retention is the name of the game that companies are playing to secure their market share.  Generally, in a competitive and saturated telecom market, as it currently is, providers gain more by keeping their existing customers, not acquiring new ones. However, it’s not that easy as it seems.

Why Customers Are Leaving and How to Stop Them

The aim to focus on retention sure sounds great, however, to accomplish this task, providers need to know, at what rate customers are switching and why.

The percentage of service subscribers who cancelled their subscriptions within a given time period is referred to as churn rate or customer churn. Naturally, for most (if not all) telecom providers, this is a crucial metric that should be regularly monitored. While for the developing markets or emerging niches growth rate is the primary measure of success, for stable and nearly saturated markets churn rate is the best indicator that shows where the company is at.

It is usually different from service to service even within the same company. For example, for the Dallas giant AT&T postpaid phone churn for 2020 was reported to hit 0.79% (thanks to their record-low fourth quarter which lowered the churn rate to 0.76%,) which is the second-best they ever achieved and overall very impressive. Now, if AT&T’s churn rate for wireless postpaid customers, which was 0.98% for 2020, is compared to the one above, it will seem that something went wrong with the service and everybody did not do a good job. Of course, the answer is a bit more complicated than that.

Customers can leave for different reasons, and it’s not always within the service provider’s area of responsibility. Sometimes, a company has little chance of influencing the decision of those people who are moving outside their service area. In this case, subscribers switch due to personal reasons or go off the grid, and that’s the part of the business that is beyond the company’s control.

But of course, for the most part, there are a lot of areas where a provider has the power to keep users loyal and keep their churn at bay. One of those, for example, is making a competitive offer to a customer with a cheaper or better alternative or minimizing customer dissatisfaction by offering superior service and support.

MaxBill supported telecom service providers for many years, accumulating vast knowledge of the industry specifics. As customer-centricity is one of the main 2021 trends and one of our key business values, our team makes sure that our partners have the means to manage their interactions with their subscribers.

Discover how MaxBill supports the multiple brands’ convergence.

A Data-driven Approach to Understanding the Churn Rate

To effectively evaluate the churn rate for the telecom company and make a well-informed decision on the proper course of action, it’s best to employ a data-based approach that consists of three steps:

• Step #1: Data Collection and Analysis

One of the benefits of automated processes, including customer management, is that all customer data can be stored securely in one system and can be accessed at any time. It is crucial to have a unified way to gather details about each consumer so that it can be analysed later. Having enough data allows for breaking down the customer base into different segments based on demographic data, behaviour, geography, revenue per period, etc. 

Modern internet service companies like Facebook or YouTube are inherently data-driven companies. They need to understand their user base intimately, how it changes over time and how to facilitate those changes. Data reporting and data-driven marketing along with short learning cycles allow those companies to continually improve key metrics week over week, month over month. This, in big part, is a driving force behind those companies’ ever-growing presence. The same principles applied to service providers ensure increased visibility of the key metrics per customer segment. This, in turn, allows for comparing churn rates dynamics in two service areas and seeing more clearly which one needs more focused attention.

Those insights are clear and actionable data: as you are doing a series of marketing campaigns, for instance, not only the results of each promotional effort can be measured per segment but different campaigns can be started if necessary based on the segment preferences.

Moreover, as companies divide the customer base into groups sharing similar characteristics, it becomes more evident how business practices impact the key metrics for a given group (segment) of the customer base.

• Step #2: Predictive analytics

Once the data is being gathered and there is operational reporting in place, insights can be expanded even further by applying self-evolving machine learning algorithms. There are plenty of methods to collect and assess data. MaxBill uses an ensemble methodology that combines several ML algorithms to achieve an impressive outcome: the results of a control run on historical data show 97% accuracy of churn prediction.

The model involves validation of the existing records of customer data against their consequent behaviour. After the tuning, a system is ready to analyse the new incoming data in real time and predict the customer’s behaviour. This way the risk of churn for each specific user can be detected and calibrated based on each new piece of information.

Our team plans on enriching the model with the additional data artefacts from a variety of segments and use them to determine which customers are at risk even faster and more precisely.

“Risky” customers, who are likely to churn, are those who require the most attention. Direct measures, be it an additional benefit or personalized offer, can be offered to the subscribers who already have one leg out the door.

Related: learn more about the role of machine learning in predicting customer churn.

• Step #3: Empower CSR’s with Data

Sophisticated analytics and the ability to pinpoint the risks for each consumer sound great, but they don’t bring any value unless the company empowers CSR’s to act upon it. The customer service representative can use the report created by the system and give extra attention to the risky customers. The analytical tool that we showcased in the previous steps can notify the CSR about any individual who needs additional attention. A customer services representative then can view the available information and take data-driven actions to engage with the consumer.

Actions to reduce the risk of customer churn often include:

• enhanced onboarding;
• personalized offers;
• discounts and credits;
• faster technical support.

CSRs can choose from the above and pick the right course of action with the support of hard data.

As industry practice shows — data and machine learning are not here to replace humans. The best results are achieved when human expertise is combined with computer-generated data.

Conclusion

There is no denying that it is hard to be a player in this game, especially when you are having a strong competition. Telecommunication providers in the UK, the US and the EU are facing the challenges of market saturation. Even though new technologies evolve constantly, opening up opportunities for adding new services like 5G connectivity, IoT support and others, telecom businesses would always benefit from improving customer retention, no matter the size and specifics of the business.

The pivotal metric to keep track of is customer churn. Knowing the rate at which a company is losing customers, observe how it changes over time, and impacts different customer segments, can give the idea of what a company needs to improve in order to retain its user base. A CRM powered with analytic and machine learning tools that supports the decision-making process of a dedicated team of CSR’s can help determine the risk of churn for each customer in real time and is a vital aid in keeping the user base loyal.

The synergy of customer service representatives and smart technology is the future of customer service. It’s that simple.