2 December 2016

Proprietary and Confidential

Measuring Progress:

Pursuing Transparency for Scale

2 December 2016

Initial observations from mapping analytics for FI

Page 2

Improving transparency, especially at national level, w/ more extensive data for a few select markets

• Ability to compare high-level progress across markets to assess impact of overall initiative portfolios

• Some opportunity to overlay national socio-economic data to consider structural drivers of FI

• Localized data available for some “headline” markets in FI, but are somewhat outliers given adoption

Granularity is resource intensive, but needed at the operational level, requires a new approach

• Localized pain points and segmentation difficult to assess, makes pre-implementation much more risky

• Lacking these data, it can be difficult to contextualize pilots when then looking to scale more broadly

• Digitization offers efficiencies for data collection; complexity calls for collaboration and standardization

Allows stakeholders to focus on bottom-up innovation, shift resources to targeted implementations

• Funders facilitate data aggregation / synthesis w/ more robust impact assessments to justify initiatives

• Stakeholders align market definitions, segmentation, or opportunity assessments to their propositions

• Panelists are illustrative of opportunities for a more data-driven approach to planning / operationalization

2 December 2016

High-level metrics useful in profiling financial inclusion gaps

Page 3

Sessions over the past couple of days have highlighted the need for and benefits from

financial inclusion, and we now take a look at the potential role of measurement for

identifying key gaps to address, developing new initiatives, and optimizing overall impact

7%

22%

0%

25%

50%

75%

100%

Adults Did NotReceiveWages

As Cash WithdrewAll

KeptBalance

Savings at FI

Low & Middle Income Markets

40%47%

0%

25%

50%

75%

100%

Adults Did NotReceiveWages

As Cash WithdrewAll

KeptBalance

Savings at FI

High Income Markets

Irregular incomes and cash payments tilt table

against FI, so propositions will need to both push and pull funds into formal channels / accounts

Global metrics facilitate greater benchmarking, and

can provide for interesting comparative profiles, and potentially as upper layers for dashboards

Source: Mondato analysis, World Bank Global Findex

2 December 2016

Findex regional / nat’l data provides more nuanced views

Page 4

East Asia & Pacific

Developing Europe & Central Asia Latin America

& Caribbean

Middle East

South AsiaSub-Saharan

Africa

0%

25%

50%

75%

100%

Below AverageAbove Average

Accounts at a Financial

Institution by Segments…

…and Penetration by Use Case

East Asia & Pacific Developing

Europe & Central AsiaLatin America

& Caribbean

Middle East

South Asia

Sub-Saharan Africa

0%

25%

50%

75%

100%

National level segmentation allows stakeholders to track objectives, but not targeted local initiatives

Increasing use cases adds additional data cuts, yet not granular enough for tracking intra-market experiments

Source: Mondato analysis, World Bank Global Findex

Acco

un

t a

t a

FI

% U

sin

g S

erv

ice

in

Pa

st Y

ea

r

2 December 2016

GSMA reveals MM traction, but use case adoption is “narrow”

Page 5

Airtime

P2P

Bill Payment

Bulk Disbursements

Merchant Payments

Int'l Remittamces

0%

25%

50%

75%

100%

2011 2015

%TxnVolume

P2P

Bill Pymnt

Bulk Dis.Airtime

Merchant PymntInt'l Rem.

$0

$3,000

$6,000

$9,000

2012 2015

Txn Value $US Mil

Source: Mondato analysis, GSMA State of the Industry Report – Mobile Money

Limited Change Txn

Composition / Diversity

Growth Historically Driven by

P2P, just one element of FI

• GSMA has taken a leading role in financial inclusion measurement w/ mobile money focus

• Significant growth in electronic transactions over past few years, but somewhat one dimensional

• Broader set of metrics beyond MNOs required for broader context, national benchmarking

2 December 2016

Empowering the wider ecosystem to accelerate innovation

Page 6

Situational

Context

Scaled Data

Aggregation

Targeted

Facilitation

Use Case Ecosystem

Scala

bili

tyF

und

ing

Current Situation: Increasing number of data points from multiple sources, and start of time series for tracking

Opportunity: Leverage digitization in building out

wider set of data points through collaboration

To-be Desired State: Shift resources to operational

assistance / facilitate wider ecosystem innovation

Limited private sector business case

Trusted third party to align interests

Alternative KPIs unlock synergies / ROI

Development Focus

Co

ntr

ibu

tio

n P

ha

se

Consider opportunities to bring a unified data

library to the wider

ecosystem

*Top-down initiatives are considered to generally be once-off interventions reliant on continued stakeholder funding, whereas a bottom-up approach is organic and continuous, such as collaborations to aggregate device-level data

2 December 2016

Moving towards more actionable insights on a localized basis

Page 7

Increased sample sizes and making these representative at the sub-national level would

be costly but increase ability to localize use cases (e.g., at provincial level as above)

Coast Province

Mombasa County

Less defined by geographic sampling, and more about aggregation of real-time device data to

explore supply-demand dynamics / impact

As-is: Moving Beyond Nat’l Level*

*In reality, localized data are available for a few select markets that have been focus areas for mobile money (e.g., India, Kenya, Tanzania), but these tend to be one-off exceptions, and other countries

tending to be based on a sample size of approximately 2,000 respondents per market

To-be: Organic Database Aggregation

2 December 2016

Panelists provide insight into measurement and impact

Page 8

• Initiatives include small-plot agriculture insurance, education savings, and financial diaries

• Illustrative of challenges in measuring impact, and how data collection is resource intensive

• While always the case to a certain extent, consider where efficiencies are possible:

• Initial hypothesis formulation, research methodology, and validation of key learnings

• Innovative data collection channels / collaborations undertaken for the studies

• Potential areas where collaboration might be both most effective and manageable

• Possible challenges to be considered (e.g., data ownership / privacy, unstructured)

• Further impact assessments being considered if data were available