/ Offices across Asia, Africa
and Latin America
www.MicroSave.net
info@MicroSave.net

Position: International Research Specialist

MicroSave seeks an international research scientist (IRS) to lead an exciting new research project for the Bill & Melinda Gates Foundation.

The Agent Network Accelerator (ANA) project will conduct detailed analysis of 25 of the leading electronic/mobile-banking (e/m-banking) agent networks in the world. The ANA project will conduct two rounds of quantitative and qualitative surveys in eight countries to derive data to assess the design and performance of agent networks over time and in different regulatory and market environments. Based on the analysis of this data the ANA project will seek to identify the factors responsible for the success or failure of agent networks in different environments. The IRS will also lead the efforts to document and disseminate the lessons learned from this ambitious research effort.

Role & Responsibilities:

The IRS will work with the MicroSave team to implement the ANA project (outlined in more detail below). He/she will be based in Lucknow, India supported by Manoj K. Sharma, MicroSave’s Managing Director-Asia, and reporting to Graham A.N. Wright, MicroSave’s Group Managing Director.

The IRS will coordinate a team of experienced researchers based out ofMicroSave’s offices in India, Indonesia, Kenya and Uganda. He/she will also oversee the work of a data manager as well as a knowledge management specialist to disseminate the research findings.

The position will require extensive travel, ability to thrive in a rich diversity of cultures and to liaise closely with others in the project – including a team working on the EM-Banking Academy that MicroSave is establishing in Nairobi, Kenya.

Qualifications:

The IRS should:

  • Have extensive experience in quantitative survey design and management
  • Be a graduate or post graduate in social sciences and have a demonstrated track record in managing large research programmes –ideally comprising both quantitative and qualitative components (however extensive experience in quantitative research is essential).
  • Have a track record of well written publications.
  • Have experience in e/m-banking (ideally) or at least financial services (required).
  • Have experience of working in developing countries (ideally in the Indian sub-continent)
  • Have a passion for scaling and evolving e/m banking platforms/technologies to serve the poor

Posting:

Lucknow, Uttar Pradesh, India for four years with 25 days of annual leave (plus 10 days of public holidays) annual flights home.

Remuneration:

$100,000-125,000 per annum depending on experience

Interested? Contact Graham A.N. Wright on Graham@MicroSave.net

Project Background:

There is no systematic data assessing the design and performance of agent networks around the world. This makes it difficult to identify the factors responsible for the success or failure of a particular agent network and how these factors vary by deployment type (e.g. bank-led vs. MNO-led) and country context (e.g. regulatory regime, population density, branch penetration). The upshot is that branchless banking providers (and donors!) are attempting to address the most complex barrier to digital financial inclusion by applying general best practice recommendation based on anecdotal (rather than systematic) data.

MicroSave will conduct detailed assessments of agent networks in all eight FSP focus countries. These assessments will generate systematic qualitative and quantitative data across a wide range of agent networks, currently envisaged broadly as follows:

Country / # of Assessments / Likely Networks to Assess
Kenya / 5 / M-Pesa/Equity Bank/Coop Bank/KCB/KPOSB
Uganda / 1 / MTN
Tanzania / 3 / Vodafone/Tigo/Airtel
Bangladesh / 2 / BKash/Dutch-Bangla
Pakistan / 2 / Tameer/Omni
India / 5 / Airtel/FINO/Sub-K/Vodafone M-Paisa plus 1/2 more to be defined as the market develops
Nigeria / 4 / To be identified as the market develops
Indonesia / 3 / Ruma/ Andara Link network plus 1/2 to be identified as the market develops

These assessments will generate systematic qualitative and quantitative data across a wide range of agent networks. The resulting dataset will address several core questions (among others):

  • What are the factors responsible for the success (or failure) of a particular agent network?
  • What are the key metrics deployments should track to evaluate the health of their agent network?
  • What are appropriate quantitative benchmarks against each of these metrics and how do these benchmarks vary across deployment type (bank-led; mobile-led) and country context?
  • How is the health and reach of individual agent networks evolving over time (particularly in FSP focus countries) and how quickly are these networks extending into poor and rural communities?
  • What are the most successful (and innovative) technologies for building and managing a viable agent network in poor and rural communities and how can these be embedded in other deployments? Potential technologies may include liquidity management & monitoring systems, performance dashboards, GIS tools, etc.

This work will aim to test a series of hypotheses about CICO agent networks that is under development in collaboration with the Bill & Melinda Gates Foundation. The current draft is attached as Annexure 1

Overview: We plan to conduct detailed assessments of agent networks and succinct quantitative surveys in all FSP focus countries (see Annexure 2). The goal is to generate systematic qualitative and quantitative data across a wide range of agent networks in order to address several core questions (listed above). By answering these questions, the assessments will advance several objectives:

  • Help agent network managers identify their network’s strengths and weaknesses and enable them to benchmark performance against comparable networks. (The core audience for this project is agent network managers).
  • Inform the curriculum of the E/M-Banking Academy and provide a foundation to track the evolution of agent networks after participants attend the E/M-Banking Academy
  • Enable donors and policymakers to track the evolution of agent networks over time, allowing them to better target interventions to address remaining gaps/deficiencies.
  • Generate a rich, cross-country comparable dataset across several agent networks, stimulating research into the deployment- and country-level indicators that predict the success of agent networks.

Research Advisory Committee: This project requires rigorous qualitative and quantitative research. Since MicroSave, as an organization, does not have deep expertise on quantitative research and statistical analysis, we seek to hire an international research specialist to lead the agent assessment studies. We will also constitute a Research Advisory Committee (RAC) to inform the development of the research methodology, common survey questionnaires, sampling methodologies, data collection practices to ensure cross-country comparability and database design/maintenance. The RAC will also identify topics for future research and comment on the project’s overall research dissemination strategy. The RAC will comprise leading researchers and academics who are experts in quantitative and qualitative data collection and/or the design of agent networks. The Bill & Melinda Gates Foundation will hold a seat on the RAC.

MicroSave – Market-led solutions for financial services

Annexure 1
Draft Hypotheses: Building a Viable CICO Network in Poor and Rural Communities
Environmental factors: CICO agent networks are more likely to be viable in countries where (i) agents are permitted to conduct KYC on behalf of providers; (ii) agents can open accounts immediately without having to wait until physical application documents have been approved by branch officials; (iii) (where agents cannot open accounts immediately), customers can at least test the new system by conducting one deposit, transfer, and withdrawal (subject to volume limits) before their account has been activated; (iv) individual agents do not need to be approved by the regulator; and (v) more than 70% of the population lives within 20KM of a bank branch (to facilitate agent rebalancing); (vi) more than 70% of the population lives within 5KM of a paved road (to facilitate agent rebalancing); and (vii) the most prevalent remote payment mechanisms are slow, risky, and/or expensive (specific metrics TBD).
Three-staged roll-out strategy- To avoid agent attrition in the early stages of a deployment, agent network managers should take a three-phased approach to building their network:(i) recruit an adequate number of agents to support a commercial launch (300 in the case of M-PESA); (ii) temporarily redirect resources from agent recruitment to customer acquisition after launch; and (iii) once the deployment has achieved a balanced ratio of customers per agent (~600 in the case of M-PESA), grow the two in parallel. Ultimately, operators will know when they’ve found this equilibrium when fewer than 10% of users report being unable to conduct a transaction within a 3-month period indicating that users consistently have convenient access to float – because agents, in turn, conduct enough transactions to reward them for maintaining this float.
Technology: To maintain customer trust in the payment system, agents need to operate in a reliable technology environment, so it is essential that the payment platform is (i) up for at least 95% of the time; (ii) supported by clear communication to both clients and agents whenever the platform goes down; (iii) enables minimum security standards (specific standards TBD); (iv) supports authentication and non-repudiation in a simple and reliable manner.
Number of transactions per agent per day: Digital money agents may be willing to maintain adequate stocks of digital money for up to 6-months even if they are earning meager (<$5 /day) transaction commissions. However, dedicated digital money agents must process a minimum of 50 transactions per day to maintain adequate stocks of digital money for more than six months. Part-time agents running a cash-in/cash-out business alongside other retail businesses must process a minimum of 20 transactions per day to maintain adequate stocks for more than six months.
Commission structure: Commissions are crucial to the sustainability and engagement of agents, so (i) agents must earn commission off every transaction they make (including deposits), regardless of whether customers pay for that transaction, (ii) commission rates must be transparent, clearly communicated and at least $0.10 per transaction to ensure that agents do not avoid processing certain types of transactions (e.g. deposits); and (iii) commissions must be paid within at least one month of facilitating the transaction. To incentivize new customer acquisition, particularly in the early days of a deployment when transactions commissions are low, agents must be paid at least $1.00 for each new customer registration and at least 50% of this commission should be contingent upon the customer conducting a first deposit.
Training: CICO agents must be given a minimum of 4-hours of basic training at the start of their operations and refresher training at least every 6 months thereafter and sales agents. Those selling more complex savings, credit, and insurance products must be given at least 2-days of training at the start of their operations and refresher/supplementary training at least every 3-months thereafter.
Monitoring visits: Regular monitoring visits and other controlling mechanisms (such as regular reporting, frequent and surprise audits) are essential to ensure standardized customer service across the agent network – both to resolve agents’ problems and to ensure compliance with process, liquidity and marketing norms, thus ensuring a consistent customer experience. Therefore, (i) CICO merchants must be monitored at least once every two weeks in the steady-state (and at least once per week during the first three months of their operations); and (ii) sales agents must be monitored every week (and typically twice per week during the first three months of their operations). Agent network managers should also employ mystery shopping on a random sample of their agents on a quarterly basis.
Marketing support: Effective branding and promotion is essential to drive volume – both of sign-up and use. Agents must be (i) consistently branded; (ii) visibly displaying current tariff posters, and (iii) supported with above-the-line (to create awareness) and below-the-line advertising and promotions/campaigns (to drive usage).
Customer support: Effective agent networks must be supported by a proactive and rapidly reactive call center that responds to agent and customer enquiries and keeps agents and customers informed about changes in services and products. Different lines should be available to agents and to clients. These must be easily accessible / reachable by consumers and agents, with appropriately defined service levels and operational at least 8 to 12 hours a day and 5 days a week.

Annexure 2

Detailed Methodology:

Agent Assessments

We have conducted initial agent network assessments in India and these will form an initial blueprint for the methodology going forward. The assessments will combine both quantitative and qualitative questionnaires to generate data on each network’s performance indicators, liquidity management structure, technology platform, revenue sharing arrangements, commission structures, and agent selection, training, and monitoring systems (see Annexure 3 for current draft of survey instruments). The core assessments will proceed in four phases:

Phase I: Desk Research, Design of Research Plan, Questionnaires, and Interview Guides: During this phase, we will conduct in-depth desk research to understand the status and nature of agent networks in FSP focus countries. We will then work with individual agent network managers and other stakeholders to develop questionnaires that evaluate each layer in the agent network, including agents, master agents, agent network managers, trainers, monitors, bank partners, and other key informants. We will also refine the set of clear, testable, stake-in-the-ground hypotheses, developed by MicroSave in collaboration with FSP (see Annexure 2), for each agent network segment. The hypotheses will be tested on each agent network assessment. The project will also provide an opportunity to begin to learn lessons about and develop hypotheses and a clear research component to assess what makes a strong master agent or aggregator. Recognizing that the current agent hypotheses lacks clarity on master agents/agent aggregators, and that this is an important area that is likely to require additional focus, on the basis of the research, in year 2 we will assess the need/opportunities to amend or extend the research protocol to specifically address the role of master agents/agent aggregators.

The questionnaires and assessment methodology will be tested on a small sample and refined based on the survey pilot.

Phase 2: First round data collection, workshop, and report-writing: In Phase 2, the finalized questionnaires will be administered. The data collected during this phase will inform the first agent assessment report for each deployment and will include a set of recommendations for how each deployment can improve the quality and reach of its agent network. The individual agent assessments will be presented to the agent network managers in a confidential meeting at the ANM’s offices. Upon completion of the analysis of all the agent networks to be assessed in each country, we will run a country workshop to present and discuss the findings with stakeholders in that country.

Phase 3: Refine the questionnaires and data collection methodologies based on stakeholders’ comments:In Phase 3,the team will refine the questionnaires and data collection instruments to reflect feedback from stakeholders during the workshop in Phase 2.

Phase 4: Second round data collection, day analysis, workshop, and report-writing: In Phase 4, the revised questionnaires will be administered in a second round of data collection ~12-16 months after the first round. This will generate insight into how the agent network has evolved over time. We will then develop a second report analyzing factors which explain this evolution and identifying areas for improvement. The individual agent assessments will be presented to the agent network managers in a confidential meeting at the ANM’s offices. Upon completion of the analysis of all the agent networks to be assessed in each country, we will run a country workshop to present and discuss the findings with stakeholders in that country compared to other countries using anonymised data.

In countries where there are more than 2 agent networks to be assessed (e.g. Kenya, Tanzania and India and possibly in the longer run Indonesia or Nigeria), MicroSave will look to engage and train a local market research agency to conduct the basic quantitative data collection across (hopefully) a larger number of agents. This will require very careful training, supervision and quality control, but can be conducted alongside the more in-depth qualitative analysis conducted by the MicroSave team, and thus marginally reduce the costs of the assessment while at the same time greatly enriching the data and analysis.

Phase 5: Synthesis, Analysis and Report-writing: After we have completed the two rounds of assessments for a number of deployments, we will analyze the findings of the assessments to check if the hypotheses stand supported or not. We then will prepare a range of synthesis documents which will (i) assess the state of play of agent networks in a given country and their evolution over time; (ii) compare/contrast different approaches to agent network management (e.g. liquidity management; agent training; etc.); and (iii) offer detailed technical guidance on select topics (e.g. how to monitor agents).