Blogback

Small data analytics in Kenya: A Case Study on understanding and driving usage

Jul 19, 2016
For the Kenya Post Office Savings Bank (KPOSB) and the WSBI Programme, a really powerful tool for better understanding the client journey was the analytical work we collaboratively started late 2014 with the aim to understanding reasons and finding solutions for people not using their account after a first initial deposit had been made. Account dormancy is an industry-wide problem. It turned out to be bigger than expected, although at most WSBI's Programme partner banks, inactivity is less pronounced compared to the industry average rates for formal financial institutions[1]. Account inactivity rates
2014Industry average inactivityWSBI Program partner inactivity
El Salvador


58%43%
Indonesia62%61%
Kenya54%45%
Morocco46%64%

Tanzania
23%[2]63%
Ugandan/a24%
A dormant account is a lost opportunity. Account dormancy or inactivity means different things to different groups of people. Irregular use and periods of inactivity are common patterns. Some clients use the account to save for a purpose, others may only access it in an emergency, and farmers save and withdraw alongside their crop cycles. It is crucial to provide multiple use options from the start and data analytics can help to classify usage and to tailor marketing messages. Therefore WSBI and KPOSB wanted to analyses and test how best to re-engage with the customer through an in-house built solution for data analytics and messaging. Some of the questions needing answers: What is a typical time gap between the first and the second transaction? Are our customer's savings and transaction patterns changing if nudged by messaging? Is there any hope that some of the accounts that have only seen an opening deposit may come live again? How frequent do customers require 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 money product, these benchmarks fell far short of expectations. From May to August 2015, KPOSB sent inactive customers of the 3,500 sample "reengagement messages" to test whether this could increase their activity levels. Whilst some customers increased their savings balances, the messaging failed to produce the expected boost in activity rates. What we found was that in order for the messaging to work, the timespans between the first client contact and the following message shouldn't be too long (50 days instead of three months). This is because the majority of active users who transacted multiple times (more than 70%), would do their second transaction within 50 days of the first contacts made. Small-data analytics work best for datasets not going beyond 300,000 transactions and are a simple segmentation tool for understanding people's transaction cycles and for identifying non-use. It is however not a silver bullet for massive take up in account activity. An in-house-built solution stands and falls with the insights into how macros can be built around the customer journey and how (seasonal) transaction gaps could be interpreted. Tracking and interpreting larger client numbers requires big data analytics, i.e. proper business intelligence tools. Subsequently nudging customer behaviour requires repeated client engagement and messaging. Therefore it looks more like an investment than an operating cost with an immediate pay-back. Investment costs into paying specialist FinTech firms for doing the interactive messaging business need to be looked at in this light. Purchasing segmentation tools externally is costly, even for mid-sized banks targeting the unserved poor. Donor support can really help here to operationalise client centricity. As we look for comprehensive solutions to emerging dormancy, doing qualitative research with customers before engaging in quantitative analysis and sending out messages turns out to be absolutely essential.

About the Authors Weselina Angelow is part of WSBI (World Savings and Retail Banking Institute)'s global efforts to providing an account for everyone and making a contribution to universal financial access. She manages the WSBI Programme for making small scale savings work, a Programme run with WSBI member banks worldwide. Winnie Omondi is an Assistant Manager - Business Systems Support at Kenya Post Office Savings Bank (KPOSB) She has been actively involved with project work at Postbank with various local and international partners where she handles data extraction for onward analytics by these partners. Her focus is on deriving patterns of account usage and studying customer behavior over time. Benson Wanyoike is a graduate in Marketing with nineteen [19] years' experience in fields related to Marketing & Sales, Customer Service, Microfinance and Management of Alternative Banking Channels. He has experience in capacity building in the above areas and has interacted with various stakeholders in managing customer relationships. He participated in the process of developing a customer service strategy for Postbank and the transformation of Postbank Kenya. _________________________________________________________________ [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.

Your comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.