During our annual DataHack4FI competition, we bring together data enthusiasts and emerging technology companies. We challenge them to extract value from data and digital technology in a way that can lead to greater financial and economic inclusion in Africa. Evolving from its origins as a short-term competition for data scientists in 2016, the DataHack4FI now brings together – over a number of months – a variety of stakeholders and participants in seven countries: 228 individuals in 2019. The participants include young data scientists who enrol in online coursework, data mentors who assist with the development of new solutions, and the emerging technology companies whose data-driven solutions are making an impact in the market.
We’ve seen numerous digital financial innovations over our three seasons of DataHack4FI. They have ranged from digitised savings and credit to innovative ways of capturing previously invisible economic activities, such as those of informal micro-entrepreneurs or small-holder farmers. The solutions pitched during Season 3 were no exception and continued the upward trend in innovation. What stood out were the notable ways in which the start-ups innovatively used data for more rapid scaling and go-to-market strategies. In this article, we describe four of these strategies, as applied by the innovators in DataHack4FI Season 3.
1. Non-financial start-ups solving challenges that underlie financial inclusion
Financial exclusion is a multi-faceted problem that is influenced by a variety of factors. Financial illiteracy, for example, means that many financial services are too complex or intimidating for some people to use. A lack of appropriate documentation (including formal records of income and economic activities) is a barrier for individuals who aim to access financial services. And, of course, there is the challenge that many financial products are simply not seen as providing sufficient value for their proposed target market. After all, individuals don’t want a financial product in and of itself – they are seeking to run and grow their businesses, send their children to school or live comfortably in retirement. Financial products are simply the tools they use to achieve these objectives.
A number of non-financial tech start-ups used their unique data assets to solve some core problems that underly financial exclusion. Examples include: the Season 3 winning solution – Botlhale AI’s polyglot chatbot Naledi – who assists South Africans to engage in their home language with financial service providers, and Fineazy’s chatbot that assists low-income customers in Ghana in their learning journey to becoming more financially literate. Solutions like that of Bureau Vente Limited and Dytech are increasing transparency and access to information in agricultural value chains – Bureau Vente through using a blockchain solution to track sustainably farmed coffee in Uganda and Dytech through using drones to predict agricultural yields in Zambia. Although the core solution increases transparency and access to key information for running agricultural businesses, the knock-on effect is that farmers are able to better save and access credit for their various financial needs.
2. Data partnerships for more holistic customer views
Access to data is crucial in understanding customers’ lives and designing products for them, but a single data source can only ever show a sliver of an individual’s life or of a business’s performance. Data partnerships are required to achieve a more holistic view. In considering different data types for a more holistic view, alternative data sources are particularly helpful for expanding financial inclusion due to their ability to make the invisible visible. Such data partnerships have the added advantage of allowing start-ups to more rapidly access large quantities of data on their customers without having to run for months or years to generate these records themselves.
The participants to DataHack4FI Season 3 provided a number of such data partnerships. For example, Kinda Credit in Kenya links to the national ID database (the Integrated Population Registration System) for access to key demographic data and reduced customer friction in sign-up. JQuicker in Rwanda partners with mobile network operators (MNOs) and government agencies (such as the Rwandan Utilities Regulatory Authority), combining data from various sources with their GPS tracking of motorcycles to generate complete views of motorcycle drivers and their transaction patterns. Extra Technologies , also in Rwanda, creates linkages between agricultural cooperatives, banks and SACCOs to generate complete transaction overviews for cooperatives and their farmers, which allows for an integrated view that combines both agricultural activities and bank transactions.
3. Aggregated models for distribution at scale
The distribution of financial products to the low-income target market is notoriously difficult. It is hampered by, among others, infrastructural challenges, low margins and challenges in marketing to an audience that doesn’t necessarily feel “financial products are for them.” To overcome this challenge, a common solution in the past has been to make use of so-called aggregators: groups of individuals, networks or existing user bases that can be reached through a central point. Well-known examples in financial services are employee groups for group life cover, or MNOs for mobile-based insurance.
Examples from the recent cohort of DataHack4FI teams include the chatbots by Botlhale AI and Fineazy. They have designed their product as a B2C offering, targeting existing financial service providers (FSPs) with their service to ultimately reach the FSP’s existing customer bases. Other start-ups target existing groups of individuals, which may be more informal in nature, to improve their functioning for their existing members. Esusu Africa in Nigeria, for example, has designed a product that increases the ease within which esusu collectors (central nodes in savings groups) can carry out their activities. Targeting these central individuals allows for them to reach the larger related savings group. And Extra Technologies, in Rwanda, targets agricultural SACCOs as aggregators for larger numbers of small-holder farmers.
4. Multi-sided platforms
The fourth and final interesting scalable model we observed among the recent DataHack4FI cohort was that of multi-sided platforms. Multi-sided platforms are a compelling model for reaching scale with data-driven solutions, given that they self-generate increasing amounts of data to be used for increasingly valuable interactions. By creating a key platform that enables a core interaction, containing a strong value-proposition for both providers and consumers of goods and services, a virtuous cycle of network effects can be set in motion. This means that the more providers that sign up, the more valuable the platform becomes for consumers, meaning more consumers sign up – which in turn leads to more providers signing up. This exponential growth of possible platform-brokered linkages can be added to the platform without any significant increase in brokerage involvement of the platform itself. Such platforms provide value in three ways. Firstly, platforms act as a marketplace for connections, allowing individuals to access more goods and services at attractive prices or to access larger markets to earn an income. Secondly, platforms aggregate valuable information (data) that can be used to design products for a variety of marketplace participants. And thirdly, platforms serve as a trusted distribution mechanism for financial services that are uniquely suited to consumers’ familiar and regular activities, meaning that financial products are likely to be more valuable and benefit from the trust that the platform enjoys.
The examples from our DataHack4FI pitches are Akello Banker in Uganda and eMsika in Zambia. Both of them are agricultural marketplaces that link small-holder farmers, agricultural extension service providers, and owners of idle farming equipment. Usalama Tech in Kenya is a marketplace for emergency response service providers that creates one centralised platform from which individuals can call the closest emergency response provider, thereby saving both parties time and money.
Collaboration is key for scale, but this may require trade-offs against rapid testing and learning
These examples describe good considerations in designing scalable digital financial services solutions. As will have become clear, the successful execution of these models does hinge on effective collaboration. For example, collaboration may be required with owners of other data assets, or with FSPs or MNOs as aggregators of target markets and holders of the required licences. Effective collaborationshould not be underestimated as collaboration can slow down the more rapid testing, learning and scaling that start-ups individually could achieve. It is encouraging to see that many of the pitching start-ups highlighted their existing partnerships, which bodes well for the scale to be achieved.
These insights are but a few of what we’ve seen from the DataHack4FI Season 3. In the next few months, we will release a number of additional outputs. We will share insights on the different data types used, the players in the ecosystem and the continuing journey of our most successful teams. If you would like to know more about our research, please reach out to Renée Hunter and if you would like to know more about the DataHack4FI, please reach out to Dumi Dube.