Skip to main content

Data can deliver big impact


For the rapidly evolving digital financial services industry, data analytics holds a lot of promise. It can produce necessary business intelligence for service providers to fine-tune product development, sharpen marketing efforts, and improve the strategy to better reach the unbanked. Big data analysis may be particularly useful to better target potential new users.

Big data is a big topic. Rarely a day passes without news of innovative applications of the data we all produce through our frequent use of technology. It is also increasingly recognized that effective analysis of large amounts of technology-generated data can support efforts promoting development. By the very nature of the business, the (Digital Financial Services) DFS industry produces a multitude of data that can be helpful in advancing its business goals and financial inclusion. Most DFS providers in Sub-Saharan Africa specifically target the previously unbanked – consumers the financial industry has previously known little about. MNOs are particularly well placed to leverage their data, producing large volumes of both call and transaction records that so far have been surprisingly underused.

To improve the understanding of existing and potential DFS users in Africa, big data projects in partnership with Mobile Network Operators (MNOs) can shed light on opportunities for new markets. Leveraging an MNOs mobile money transaction database and call-detail records, data analytics can answer questions such as: What characterizes active mobile money users? What drives inactivity? Is it possible to identify behavior patterns among customers and to use that information to stimulate better uptake of the service? Is it possible to better target potential new customers that are more likely to be active DFS users? The objective of analytics can help answer important questions such as how to improve uptake and active usage of services.



In Africa, such studies have been carried out, for example, using six months of call detail records and mobile money transactions data, together nearly one terabyte in size but this can vary depending on the size of the MNO. Users were segmented into three categories: ‘voice only’, ‘registered but inactive mobile money’, and ‘active mobile money.’ The results showed that these segments had very distinctive patterns of usage in terms of voice calls, social network structures and geographical mobility. Active DFS users, defined as customers who use digital financial services consistently at least once per month, made on average not only almost twice as many phone calls than customers not using mobile money, but these calls also lasted significantly longer.

The same pattern was true for text messages: active mobile money users sent and received the most SMS, followed by inactive mobile money users and then by non-users of DFS. Whilst active DFS users call and text their friends, families and business partners more often, they also have a much larger social network with than other users, and these contacts are geographically more spread. Non-users of DFS seem to move around much less than active users, as evidenced by a lower number of cell towers picking up their phone signal. Active DFS users moved around ten times more than voice-only users. Factors such as lifestyle and mobility between rural and urban areas may account for this.
While more research is needed to better understand these customers and determine whether regular usage of mobile money increases usage of other services, it is clear that they constitute a specific and valuable customer segment that is geographically mobile and socially well connected. For example, it is possible that some customers have small businesses for which they use their DFS accounts, which could explain both a large network of contacts and high mobility.

In one such research, the data analysis revealed a strong correlation between high usage of telecoms services and the potential to be an active, regular DFS user. The big data did not contain any socio-economic or demographic information however, and to overcome this limitation the project combined the big data analysis with classic market surveys. The surveys included age, gender, and education analysis, as well as income and key poverty index metrics.  The MNOs users represented all demographics, but tended to be more male; married; in their early 30s; high-school educated; and living in lower-middle tier housing. 

The overall research showed that many telecoms-only customers had a demographic profile similar to highly active DFS users. The team, therefore, scored all telecoms subscribers according to the extent to which users are similar to the profile of highly active DFS users, using a model that captures some dominant usage variables to predict whether a subscriber is likely to become users of mobile money. Based on this predictive model, an MNO can translate this to maps that identify areas of the market with the highest concentrations of likely adopters and to concentrate marketing efforts there.

A couple of interesting general insights have also been gained from this research. Whilst younger people are the most active users of voice services, for example, they also have the largest share of registered yet inactive DFS accounts. This suggests there is room for improvement of the services offered by DFS since these mobile-savvy younger customers are not being engaged by the current offering despite using their phones regularly for other reasons. The study also confirmed that DFS usage leads to loyalty and increased usage of same-brand services.

DECLARATION OF ORIGINALITY

I affirm that the attached work is entirely my own, except where the words or ideas of other writers are specifically acknowledged according to accepted citation conventions. This assignment has not been submitted for any other course at Robert Kennedy College or any other institution. I have revised, edited and proofread this paper. I certify that I am the author of this paper and that any assistance I received in its preparation is fully acknowledged and fully disclosed in this paper (examination). I have also cited any sources from which I used data, ideas, theories, or words, whether quoted directly or paraphrased. I further acknowledge that this paper has been prepared by myself specifically for this course.

Comments