2. contact@e-mfp.eu
What are this session’s objectives?
1 Understand the Big Data techniques in the context of financial inclusion
2
Identify what the benefits of Big Data can be for customers and providers
Learn how to put Big Data techniques in practice
2
4. contact@e-mfp.eu
Who are this session’s speakers?
Alexis Label, CEO, OpenCBS
Data collection systems and apps.
solutions, digitalization of appraisal
process
Etienne Mottet, Innovation Analyst
at Business and Finance Consulting
Should we mine the big data in
microfinance? Introduction with
farming case studies
Yasser El Jasouli Sidi, Fonder, MFI
Insight Analytics
Data analytics in Microfinance how
does it work, practical example of
Credit scoring.
Simon Priollaud, Digital Financial
Services Consultant at Inbox
Practical experience of projects in
Africa on Big Data, results and
lessons learned
6. contact@e-mfp.eu
Case comparison
2 Farming
activities
1 Challenge
Get the best yield and
profit from their fields
Tylek from Tuyuk Village,
Kyrgyzstan
3 Ha of wheat
2 Ha of parleys
30 livestock head
10,416 Ha
All cultures
High
Mechanization
How is Data being used to address this challenge?
7. contact@e-mfp.eu
III.
Decisions
I.
Data
Collection
II.
Analysis
- Investment in sensors, GPS,
tractor fleet guidance tools
- Big Data agro analysis software
- Live tracking of input and
tractors
- Precision mapping of yield and
other indicators
- Invest in better intelligence
- Tractor fleet management optimization
- Configure input usage automations
- Tractor auto-control
Benefits: 15% input saving, 20% income increase, better cost control, better soil management
8. contact@e-mfp.eu
Tylek from Tuyuk Village
- Potential for beetroot crop in the region
- Nutrients in soil suitable for growing
beetroot
- Agro expert scoring for the application
- Automated crosscheck with online credit
bureau
- Data analysis & statistical scoring
development
- Consider a new type of crop
- Development of specific
beetroot agro product
- Tylek applies for beetroot loan
- Disbursement decision
Started in 2008. 3000 farmers growing sugar beets.
Factory at max capacity and 2nd factory to be operational by September 2017.
III.
Decisions
I.
Data
Collection
II.
Analysis
- Sugar factory under capacity in Chui Region
- Tylek learns about beetroot opportunity
- Sensor test on field nutrient composition
- Field client information collection
- Product results collection
9. contact@e-mfp.eu
Situation Comparisons
Live connected
equipment
Credit Bureau integration
Expert scoring
XLS based analytical scoring
Digital
information?
Deeper data mining?
Tylek from Tuyuk Village,
Kyrgyzstan
Where could technology improve the process?
III.
Decisions
I.
Data
Collection
II.
Analysis
Management decision
Configure automation
(AI)
Agro Big Data solution
Mapping representation
Expertise & score based decision
One-time soil analysis
Client info collection
Word of mouth + farmer gatherings
Knowledge of context
Tablet info
collection?
10. contact@e-mfp.eu
Follow the digital footprint?
Or nurture strong knowledge of
context
Embrace the internet of things?
Or use simple tech smartly
Mine Big Data?
Or smartly leverage existing data
What matters in our context of operation?
11. 13
Big Data or Small Data?
— Should we mine the Big Data in microfinance?
Maybe…
— But let’s pick small data first!
Thank you!
13. contact@e-mfp.eu
A free CBS with payable add-ons and services
Additional modules & custom developments
Implementation
Technical Support & Software maintenance
Training of users
100 free users and 20 paying clients
A team of 16 in Bishkek and Hong Kong
More than 10 year experience in Microfinance
OpenCBS introduction
22. contact@e-mfp.eu
4.0 Credit Scoring Use Case
Situation
A small loans provided with wide network was making manual decision in face to face
meeting. Decisions were made manually under wide guidelines. Customers would take
multiple loans each year, often with 2 loans running in parallel.
What we did
• We introduced customer management system and behavioural score. On each cycle
• point a score and maximum limit was calculated and 3 possible recommended new loans
• made for those customer which were eligible by the system rules.
The result
• Client facing staff appreciated the support and guidelines. Benefits were seen in both;
• Increased sales where sales staff too conservative reduced losses to higher risk customers
• whose relationship with the staff made it difficult to say no
23. contact@e-mfp.eu
4.1. Risk Mitigation
Credit Assessment can be
done before lending out
loans using Financial Data
and Alternative Data and
such as:
• Demographic Data
• Social Data
• Mobile Data
24. contact@e-mfp.eu
4.2. Value of Credit Scoring
Risk Assessment Product Offer
Score Product Name
Overall Risk Suggested Loan Amount
Default Probability Suggested Collateral
Odds Annuity
25. contact@e-mfp.eu
4.3 Impact of Credit Scoring
• Credit Scoring Tools assists in
cleaning the assets by eliminating
borrowers that are not credit worthy
and may effect the portfolio
delinquency and default probability.
• Fewer calculations are needed for
performing data search
26. contact@e-mfp.eu
4.4 KPIs
• A decrease in the loan
turnaround time from 72 to 6
hours
• An increase in average loan
officer caseload of 134
percent
29. contact@e-mfp.eu
In the three last years
• More than 35 projects in 5 years (3 > 1.8 M.
EUR in DFS)
• Commercial segmentation in more than 20
countries in Europe, Africa & Asia
• Largest client has 22 million of clients
Our track record in Africa2. Inbox’s experience
30. contact@e-mfp.eu
3. Results - Definition of some segments
Know my clients
Understand my
clients
Better serve my
clients
Think about the
next move…
1. Audit MIS &
environment
2. Identify my
segments
3. USE the
segmentation
objectivesSteps
31. contact@e-mfp.eu
3. Results - Definition
of some segments
Youth
(<18 ans)
(3 segments)
Inactive
(9 segments)
Low income
people
(3 segments)
Clients without
savings account
(6 segments)
Clients with checking
account
(7 segments)
Amountcreditedoverthelast12months
Overall balance
33. contact@e-mfp.eu
3. Results - Definition
of some segments
Large companies
SME
Microenterprise
VIP Clients
« Working class »
Mass market clients
Low income clients
Commercial
Segmentation
My environment
tomorrow
(hopefully…)
My
environment
today
34. contact@e-mfp.eu
4. Some advices
1. Do not copy-paste : what you need has to be tailored.
2. Take your time and assess the data you have in your MIS, you most probably already
have all the data you need.
3. Do not underestimate your MIS : segmentation could be integrated in most MIS.
4. Segmentation is a useless tool if you do not use it continually and update it regularly.