Mobile money activity rates: How other industries have boosted usage

This is the third of four blogs aimed at understanding the causes of low user activity rates in the mobile money industry and potential strategies on how to improve mobile money activity. The previous blog explored past mobile money account activation strategies within the industry. This blog focuses on successful customer activation strategies in other industries. The next blog will summarise findings from the previous three blogs and issue recommendations.

Several organisations outside of the mobile money industry have pursued a range of strategies to improve customer activity for their products and services. However, no single strategy – whether successfully tried or not – can be reused identically in a mobile money setting, let alone by any specific provider, due to differences such as geography or resources at hand. This blog focuses on organisations with attributes comparable to the mobile money industry, whether it is the type of activity (banking, fintech) or the tools and assets they have at hand (mainly data). These examples can offer ideas which mobile money providers can pick and experiment with to increase customer activity.

Churn prediction

Several data-driven organisations have been using churn prediction methods to improve customer retention and engagement, boosting activity rates in the process. Commonly achieved through the use of data science, churn prediction is defined as the detection of customers who “are at high risk of leaving a company or cancelling a subscription to a service, based on their behaviour with [the associated service] or product”. Companies that pursue this strategy can build models based on past behavioural patterns from customers who eventually churned (i.e., stopped using a product or service), to detect current customers with such behaviours before they stop using their products and services. However, predicting churn is only useful if corrective actions are taken to then prevent identified users from stopping to use one’s service.

Paytm, a leading India-based fintech, uses churn prediction as one of its key customer retention and growth strategies. It does so both internally, including through its Growth Marketing team, and through specialised external providers. For example, Paytm has partnered with marketing company Airship to use predictive artificial intelligence to identify users at risk of churning, and then re-engaging them through relevant offers. Together with other solutions, such as segmentation and personalised messaging, this strategy helped to increase the number of Paytm’s premium subscribers by 81% per cent year-on-year in 2020. Though Paytm operates in a very large market and is backed by sizable investments, mobile money providers can also use a similar approach to identify accounts at risk of becoming inactive before they do so, and deploy preventive measures. This approach could be deployed for customers who use mobile money, even if irregularly, but might have reduced their use for any given reason.


Hyperpersonalisation is the process of “creating custom and targeted experiences through the use of data, analytics, AI, and automation”. This can translate into tailored communication, product design or even pricing, based on each individual customer’s behaviour and preferences, whether permanent or temporary. In the financial sector, hyper-personalisation can occur on the back of transactional data, customer feedback or even individual interactions. For example, in Hong Kong, HSBC has been using the spending and saving behaviour of its customers to provide personalised advice to their customers.

The wealth of data collected by mobile money providers may not draw the financial profile of each of their account holders as accurately as banks do for their customers. This is due to the use of a smaller set of use cases in the mobile money industry. However, transactional data may still offer opportunities to personalise communications, services or promotions, especially if conducted in conjunction with other strategies, in order to reactivate inactive accounts or prevent inactivity. Importantly, this approach is relevant as reasons for low or no activity can vary greatly among users. For example, one individual may use mobile money infrequently as they only receive occasional remittances or scheduled cash transfers, while another may have stopped using the service due to alternatives perceived to be better or cheaper.

Mass experimentation

Mobile money providers can help boost activity by both improving their existing services and widening the range of services offered to their users. But how should providers choose the right services with the right design? The rapid testing of many features, whether interface-related or fully-fledged services, can sometimes provide the answer, especially for companies with sizable customer bases where large-scale testing is possible. For example, companies such as or Expedia typically run thousands of experiments simultaneously. Assessing their efficiency through A/B testing and relevant key performance indicators, these companies have been able to make incremental improvements to their product design, allowing them to acquire and keep millions of customers since their launch. As mainly digital companies with large account bases, mobile money providers with the right human and technical resources could run numerous tests simultaneously, gradually refining their product suite and improving customer stickiness.

Such strategies have also been adopted in low- and middle-income country (LMIC) contexts. For example, the Singapore-based super-app Grab (which operates in numerous Southeast Asian markets) launched an in-house Experimentation Platform. This has enabled the company to seamlessly manage multiple experiments concurrently, preventing interference between experiments, optimising sampling and improving analysis (e.g., through real-time dashboards).

Data-driven security

Account inactivity by certain users could at times be the result of fraud, attempted fraud, or the fear of being targeted by fraudsters. Mobile money providers already recognised this challenge and have implemented strategies to counter this – as evidenced by numerous providers having achieved the GSMA’s Mobile Money Certification. Nonetheless, over-cautious systems may at times block legitimate transactions, causing user dissatisfaction, and underlining the importance of accurate detection. Fraud attempts can be targeted through a multi-layered and data-driven approach. This can include machine-learning models, alternative data sources (e.g., geospatial data, phone usage), or automation.

Chinese digital payment giant Alipay has designed an AI-powered risk engine called AlphaRisk, both for its own use and as a security solution for other organisations. In addition to detecting fraud attempts and taking countermeasures, the system can adjust the risk profiles of subscribers based on past fraud attempts, and its “Auto Pilot” mode helps apply the best risk management strategy in each situation, drawing from more than 200 algorithms. As of 2020, the solution has helped Alipay keep its fraud loss rate to under 0.64 in 10 million, meaning that for every USD 10 million, losses through fraud are worth less than USD 0.64. AlphaRisk has already been used within the mobile money industry by GCash in the Philippines, which adopted the service in 2020. The system has improved GCash’s security (e.g., promotion abuse fell to less than one per cent), even leading to the launch of a money-back programme in case of unauthorised transactions.


While this method may not suit every market or segment, mainly due to relatively lower smartphone penetration levels and the prevalence of USSD or STK use, gamification strategies can be used to improve customer engagement and activity in the mobile money industry. Defined as “the practice of making activities more like games in order to make them more interesting or enjoyable”, it has been used as a successful strategy for user retention in the app market, including for fintechs such as neo-banks and payment platforms. For example, Revolut – a major UK-based fintech – has been rewarding customers with points when they use its services, which can be used to enter prize draws. Other gamification tactics include helping users to track their progress towards one or more goals – whether rewarded or not – in a visual manner, for example by using progress bars/gauges. Often used by eLearning (e.g., Duolingo) or fintech (e.g., Starling Bank and Revolut) apps to set and track learning, budget or savings targets, they can be adapted by money services available on apps to improve user retention, especially with growing smartphone penetration rates in LMICs.

Among the ideas explored, no single strategy can be adopted by any given mobile money provider in an identical manner as done elsewhere. Beyond market and product specificities, Important variables such as cost, technical and human resources must be considered when deciding how to tackle the challenge of low account activity or dormancy. For example, providers’ available data infrastructure should be suitable to make data-reliant initiatives possible, as many strategies described in this blog are. Ultimately, mobile money providers should consider strategies to boost usage that can account for the characteristics of the markets they serve.

If you’d like to learn more about the barriers to regular mobile money use, check out the first blog and the second blog in this series.