Engaging DFS customers through personalized targeting and segmenting the most promising client base (the case of Juntos)

Author: Oleksiy Anokhin (based on original case authored by Minakshi Ramji)

Smartphone users are overwhelmed with access to information through multiple channels – websites, social media, and messages. As a result, people can become inured to numerous everyday messages, which they may perceive as spam. How can a provider approach their customers using these digital channels without irritating them? The case of Juntos, a Silicon Valley technology company that partnered with DFS providers, demonstrates how to reach out to DFS customers through personalized messages in a way that helps to increase their activity and identify a potentially more active customer base.

Personalized outreach can be an important potential solution for DFS providers, who try to build their business as responsible finance actors. More direct and personal two-way communication can help develop mutual trust and respect, allowing customers to feel as if they’re being treated fairly. Moreover, two-way communication helps customers feel included and that their opinions are valued, which can help build trust in the product and provider. Finally, such digital outreach can be combined with other more traditional methods, such as ATL (Above The Line) and BTL (Below The Line) marketing campaigns and generic messages for more effective work with existing and potential customers.

Globally, many DFS providers experience high inactivity and low engagement. This discourages providers, whose investments may not be seeing sufficient financial return and whose customers may have access to services without sufficiently realizing the benefits of those. Juntos offers a solution to this problem by using personalized customer engagement messages based on data-driven segmentation strategies that deliver quantified results. Good data underpin this approach. First, Juntos conducts ethnographic research to better understand customers in the market. Engagements are always informed by quantitative data provided by the DFS partner, qualitative behavioral research done in-country and from learnings drawn from global experience. Having developed an initial understanding of the end user, Juntos conducts a series of randomized control trials (RCTs) prior to full product launch. These controlled experiments are designed to test content, message timing or delivery patterns, and to identify the most effective approach to customer engagement. To begin, messages are delivered to users, and users can reply to those messages. This develops the required trust relationship. More importantly, those responses are received by an automated Juntos ‘chatbot’ that analyzes the results according to three KPIs:

  • Engagement Rates: What percentage of users replied to the chatbot? How often did they reply?
  • Content of Replies: What did the responses say? What information did they share or request?
  • Transactional Behavior: Did transactional behavior change after receiving messages for one week? One month? Two months?

These experiments enable Juntos to understand which inactive clients became active because of Juntos message outreach, and to understand which messages enabled higher, more consistent activity. For example, a control message is sent to a randomly selected group of users, such as, “You can use your account to send money home!” Others might draw from service data to include the customer’s name: “Hi John, did you know that you can use your account to send money home?” Other data may be incorporated within the message: “You last used your account 20 days ago, where would you like to send money today?” These are merely examples, but they show how a generic message compares with a personalized message with a time sensitive prompt. Juntos’ baseline ethnographic data improve its qualitative understanding of customers, helping build the hypothesis around which messages are likely to resonate, then putting those messages to statistical test. The first question is whether the test messages yield statistically better results compared with the generic control message. When the answer is “yes,” it is important to dive one step deeper and ask about the respondent and surveying across segments such as rural or urban; male or female; income range; and usage patterns, merging this information with ethnographic data on consumer sentiment. By testing a wide variety of messages, Juntos is able to segment user groups according to messages that show statistical improvements in usage over time. This means that high engagement messages can be crafted for everyone from rural women, to young men, to high-income urbanites. The Juntos approach is tailored for each context and is continuously tuned to nimbly accommodate customers who change their interactions over time.

The success of Juntos’ case demonstrates that a data driven approach can be one of the options of working with customers for any DFS provider. Using various methods of data collection leads to more accurate and targeted customer outreach. Moreover, additional analysis of clients’ feedback and involvement helps to improve the initial approach and segmenting of customers more precisely in future. Finally, personalized engagement might help to build mutually beneficial and transparent relations, based on respect and trust between providers and their clients, positioning the former as responsible finance actors.

Adapted from a case study (prepared by Minakshi Ramji) presented in the Data Analytics and Digital Financial Services Handbook (June, 2017), this post was authored by Oleksiy Anokhin, IFC-Mastercard Foundation Partnership for Financial Inclusion, for the Responsible Finance Forum Blog July, 2018.