How E-Commerce Is Transforming the Idea of the Rural Agent

The number of agents in the mobile money industry has increased dramatically, doubling from 3.2 million to 6.6 million between 2015 and 2018. However, this agent network growth has come with disappointingly modest gains in agent and account activity: the share of agents active in the past 30 days went from 51 to 57 percent, while the share of individual accounts active in the past 90 days went from 33 to 35 percent during the same period, GSMA reports.

Research from CGAP and others helps explain this counter-intuitive trend. It shows a strong correlation between how close agents are to clients and those clients’ uptake and use of mobile money accounts. The problem is that despite the growth in agent numbers, this growth tends to be clustered in urban areas, leaving a large share of the population in rural areas without access to agents. This contributes to their low adoption and use of mobile money services. For example, a 2018 CGAP assessment in Tanzania revealed that despite the fast growth of the country’s agent network, only 44 percent of the population was covered. Most of those not covered lived in rural areas.

What’s stopping mobile money agents from reaching deeper into rural areas, where many low-income people live? Are recent innovations among mobile money providers going to solve these challenges? And do new players in digital financial markets like e-commerce companies have something to teach other providers about building rural agent networks?

CGAP is reviewing the diverse agent network experience around the world in partnership with digital financial services (DFS) providers, development actors, donors and regulators to answer these questions. One thing that’s becoming clear early on is that an unlikely source — e-commerce companies — are starting to provide new and useful insights on how other DFS providers can scale rural agent networks.

Leading e-commerce firms are spreading rapidly in rural areas

E-commerce doesn’t come up very often in conversations about agent networks for DFS — and, until recently, there was little reason why it should.

Previously, e-commerce companies in developing countries had focused on attracting users in big cities, where clients have internet access and are able to shop online. But lately, companies from China to Latin America have been expanding into rural areas by building agent networks. These networks are quite different from the ones built by mobile money providers in East Africa or South Asia.

A few of these networks are starting to show significant rural reach, suggesting that there is more than one way for DFS providers to establish viable last-mile agents  and make financial services more accessible and useful for low-income customers.

The rural mobile money agent business model and its constraints

To see what’s novel about e-commerce rural networks, it’s helpful to first understand the challenges faced by mobile money networks. The Bill & Melinda Gates Foundation-supported 2019 BCG study on mobile money agent economics suggests that the challenge to reach rural areas stems from agents’ underlying business model — specifically, their dependence almost exclusively on cash-in/cash-out (CICO) transactions to generate revenue, have enough liquidity and turn a profit. Because of the low population densities in rural areas, agents reach fewer clients from whom to generate transaction fees. This makes it difficult to break even, let alone make a profit.

In an attempt to build a competitive advantage, mobile money providers and regulators sometimes compound this challenge by requiring their agents to be exclusive (only represent one mobile money provider) or dedicated (only be in the mobile money business). In fact, the BCG study shows how agents that are nonexclusive and nondedicated can break even with fewer mobile money transactions because they can generate additional revenue streams by selling other goods and services or additional transactions offered by other mobile money providers.

Mobile money providers have been experimenting with promising solutions to these challenges. For instance, GSMA and MicroSave have highlighted providers’ innovative use of data analytics and artificial intelligence to solve liquidity issues. However, beyond exploring ways to make the current business model more efficient, it’s worth going a level deeper and ask whether the underlying agent business model could change.

E-commerce agent networks offer new ideas for a rural agent model

As e-commerce companies providing DFS have reached scale in urban areas, they have set their sights on expanding to rural ones, where cash still dominates. To deal with this reality they have had to establish agent networks to enable distribution of goods and some forms of CICO. In doing so, they’ve encountered many of the same constraints faced by mobile money providers. A 2017 CGAP paper documents how this story has been unfolding in China.

Some of these companies are already showing impressive gains in rural outreach. Perhaps the most mature example is Alibaba’s Rural Taobao program. Launched in 2014, Taobao has already established a rural agent network that serves 30,000 rural villages where customers can pay cash for goods and services they bought via the Taobao network. Customers without smartphones can also make cash withdrawals and conduct payment, credit, savings and insurance transactions offered by Ant Financial (Alibaba’s financial spinoff), through these agents. To get a sense of the relative scale of this coverage, note that the state-owned Agricultural Development Bank of China — one of the country’s top three banks in terms of assets and rural coverage — had 24,000 domestic branch outlets, according to its 2017 annual report.

Just like mobile money agents, e-commerce agents rely on transaction volumes to generate revenue. Whereas mobile money agents are limited to CICO fees, however, e-commerce agents can earn revenue from purchase transactions of dozens of goods and services, including DFS, that are sold on the e-commerce platform. This business model offers more agent revenue streams per client, making it easier for rural agents to become viable and offer CICO services despite low population densities.

Similar approaches aiming to have rural coverage are being developed by providers like StoreKing and Paytm Mall in India, Go-Jek and Grab in Indonesia, Jumia Pay across Africa, Oxxo in Mexico and Mercado Pago across Latin America. These providers are recruiting diverse agents — from little-known rural mom-and-pop stores to motorbike drivers — to act as roaming agents, blurring the definition of traditional agents and merchants. These new providers have stronger incentives to service rural clients given the vast product mix they can monetize, and they are mobilizing large investments to develop such agent networks as part of their broader strategies.

Learning from diverse agent network experiences

As the e-commerce example shows, there is a growing number of approaches to build more inclusive DFS agent networks . CGAP’s work on DFS distribution will compile, connect and promote consensus on lessons from the diverse agent network experience around the world. By doing so, we hope to synthesize ways to make DFS more accessible and useful to underserved and financially excluded people. We will continue to share lessons learned throughout the year.

For more information about how financial services providers, funders and policy makers can build high-quality, inclusive rural agent networks, see "Agent Networks at the Last Mile: A Guide for Digital Finance to Reach Rural Customers" (CGAP, 2019).