Understanding Customer Inactivity with Customer Data from Kenya

Account dormancy among mobile money users is an industry-wide problem. In fact, the challenge is daunting, with more than two-thirds of worldwide mobile payment accounts reported as inactive as of December 2015.

For savings accounts it is no different. At WSBI’s Programme partner banks, inactivity is less pronounced, however, than for mobile money users. Inactivity is also less pronounced at most of the partner banks compared to the industry average rates for formal financial institutions.¹

For KPOSB, the Kenya Post Office Savings Bank, a WSBI member institution, segmenting customers’ account usage data based on activity levels proved to be a powerful tool for better understanding why and how many accounts were not actively used. A dormant account is a lost opportunity, but it also gives clues into the driving forces behind inactive accounts and eventually the chance to re-engage with the customer. WSBI used text messaging as the foundation for a test KPOSB and WSBI conducted in 2015.  

Some of the questions needing answers during the analysis were:

• What is a typical time gap between a customer’s first and second transaction?
• Are our customers’ savings patterns changing?
• Is there any hope that some of the accounts that have only seen an opening deposit may become active again?
• Do customers require constant and continuous engagement?

Findings

Using a new fully enabled mobile banking product and after looking at activity within a sample of 3,500 accounts from May through August 2015 as a benchmark, we found that 79% of the accounts were inactive. Of these, 28% had zero balance.  For a new mobile banking product, these benchmarks fell far short of expectations.

From May to August 2015, KPOSB sent inactive customers “reengagement messages” to test whether this could increase their activity levels. While some customers increased their savings balances, the messaging failed to produce the expected boost in activity rates. Why didn’t this experiment work?

• Messaging needs to happen soon after the first contact with the customer has been established, and then repeatedly. Based on the assumption that the majority of active users were waiting less than 50 days to make a second transaction, nudging inactive users should happen within the same time frame.
• Periods between messages should be close together and should not exceed the number of days between the first and second transaction as observed for more active users. Messages need to be highly tailored, which can be difficult and expensive to achieve. KPOSB designed and tested messages based on a segmentation model which was built in-house by experienced system developers from the bank. While such a solution can deliver powerful insights into client transaction behavior, it reaches its limits with high transaction volumes. Further, it does not provide the same level of detail which expensive business intelligence tools would deliver. The same applies for reengagement solutions where business intelligence tools can help define more precise timespans for sending out the messages.

Tackling account inactivity requires both data analytics and engagement messaging. Both methods should be seen as long-term investments rather than an operating cost with an immediate return. Most importantly, however, is doing the qualitative work with customers before kicking off a reengagement campaign. Qualitative face-to-face client feedback is the tried and tested way to understanding what drives customers to use a product in the first place. Once we are armed with that information, we can start to whittle down that high inactivity rate.

1. Activity for mobile money users is measured for at least 1 transaction within 90 days. Activity for savings account users at WSBI partner banks is measured for at least 1 transaction within 180 days. Industry average calculations: Worldbank Findex Estimate of Total Adults with an Account at a FFI - formal financial institution (total / depositing) and IMF FAS Estimate of Total Accounts, S. Peachey Proxy Method 2015.

2. Tanzanian Banks were recently required to clean-out dormant accounts; WSBI partner bank data is based on figures prior to clean-out.