Churn models to help SA's mobile telcos build loyalty
Is predictive modelling key to understanding customer loyalty? This is the question posed by the South African marketing insight services firm, Knowledge Factory, which argues that predictive modelling techniques are the key to understanding what drives customer loyalty, particularly in the mobile phone sector where number portability has just been introduced in the country.
The introduction of Mobile Number Portability (MNP), which enables South African mobile phone service customers to retain their phone number when changing mobile service provider, is expected to test customer loyalty and trigger far higher levels of churn, as it did in other countries that have already introduced number portability.
Attention to retention But some analysts believe that more than one-quarter of consumers (a predicted 27%) will change networks within one year of the service becoming available, and such predictions have ensured that customer retention is now firmly at the top of the agenda for the country's mobile operators.
"MNP will definitely have an impact and create volatility in the telecoms market," warned Suben Moodley, client services manager for the Knowledge Factory. "But it will simply be an expansion or increase of existing churn propensities."
Moodley acknowledges that, without the artificial hook of losing their number, customers will feel less trapped, but also argues that it will be existing grievances that prompt them to switch providers: "The higher rates of churn will predominantly be from customers with existing complaints or reasons for wanting to move, simply taking advantage of MNP."
What drives churn? Although particularly rife in fiercely competitive industries such as telecoms, customer churn is a significant problem that spares very few companies. The mobile communications industry in particular is largely product driven and price conscious - two traditionally strong reasons for runaway customer churn. But Moodley suggests that the main driver of churn may actually be service delivery, both in terms of customer service and network performance.
"Although it might appear otherwise, with price conscious customers clamouring for the latest and greatest handsets, these reasons tend to cancel themselves out in the age of hyper-consumerism," said Moodley. "And products and service packages are quickly replicated by the competition anyway, so there is no real, enduring advantage from either."
Churn patterns To help companies effectively manage their churn rate, Knowledge Factory uses predictive modelling techniques based on empirical evidence and analytics to help find churn patterns. By combining a variety of mathematical techniques, including artificial neural networks, statistical regression, and decision trees, the company determines the propensity of any individual customer to cease doing business with a company within a given time period.
According to Moodley, the results are often surprising: "You might be under the impression that customers want the latest phones or trendier outlets, but - for example - a large number of those leaving the network may have experienced a considerable number of dropped calls in the past six months."
Data analysis needed Analytical customer management strategies can certainly be helpful for mobile phone network operators because they have unusually rich transactional data, which allows very specific patterns and results to be identified.
"Many operators have recognised this and made significant investments in churn modelling software," explained Moodley, "But many still struggle to get maximum value from their data."
For example, Moodley cites using a network's detailed transactional data to up-sell and cross-sell customers. Apart from reducing churn, this is one of many ways that mobile operators could be using their data to drive greater loyalty and reduce the risk of the churn that will inevitably result from number portability.
"In the end, MNP represents a tremendous opportunity for the network operators," concluded Moodley. "Armed with customer churn modelling techniques, operators can not only ensure they mitigate churn but also take advantage of the volatility for acquisition, making sure that those customers who do move are moving towards them rather than away from them."