FINCA International

1
FINCA International

• Enhancing Customer Research Personal Digital
  Assistants

2

FINCA Internationals Outreach
3
FINCAs Approach to Customer Research

• Inform social and financial priorities
• Engage key stakeholders
• Create efficiencies through tools, methods and
  systems
• Ensure accurate analyses and timely reporting

4
Customer Research Tools
Outcomes Administration
Loyalty/ Satisfaction

• 8 questions
• Ten minutes or less to conduct

• Measure client experience
• Determine loyalty drivers

• 8 questions
• Ten minutes or less to conduct

Retention

• Identify reasons clients leave or are planning to
  leave

• Measure depth of outreach/poverty levels
• Can show change over time
• Segment clients into demographic/economic
  categories

Standard of Living

• 16 questions
• Fifteen minutes to conduct

• Comprehensive measure of social financial
  well-being
• Cross-analysis of 8 content areas assets,
  expenditures, business, loan, basic needs,
  loyalty/satisfaction, and exit.

FINCAs Client Assessment Tool (FCAT)

• 120 questions
• Thirty to sixty minutes to conduct

5
FINCA's use of PDAs

• Since 2004 PDAs used for FCAT data
• 15,000 interviews conducted
• Used in more than 25 countries
• Surveyors trained on technology in 1-2 days

6
Considerations for Using PDAs

• Increases speed of data collection
• Easy to load data into a computer
• Limits data entry time, costs and errors
• Roughly 100 for each PDA (retail)
• Surveyors may be adverse to technology
• Need to be careful of theft
• Backup method requisite

7
Resources Hardware

• The following hardware supports FINCAs Customer
  Research process

8
Software on the Laptop

• Palm Desktop
• Used daily
• Allows data to be copied (sync) from Palm Z22 to
  Laptop
• Pendragon Forms 5.1 Manager
• Used daily for sync
• Stores the survey forms and lookup tables and data

9
Resources Software on the PDA

• The following software on the Palm Pilot supports
  the FINCA survey process

10
Data Warehousing

• Centralized location of social performance client
  data
• Users can pull nearly infinite number of
  indicators for analysis at global, regional or
  local level
• Improved querying capabilities allows for quick
  access to nearly 19,000 client records
• The Research DW cuts reporting time by 50 and
  reduces errors
• Allows for normalized trend analysis from 2003 to
  2009

11
Example Network Food Security by Product
12
Questions?
Pros Mugaga FINCA Uganda Client
Ms. Mugaga, a mother of four, has been a FINCA
client for four years. Her inventory has doubled
and she is able to keep her children in school.
13
FINCA International

• Demo - Enhancing Customer Research Personal
  Digital Assistants

14
Palm Pilot Parts Pieces
Reference Palm Z22 Read this First
15
Palm Pilot Parts Pieces
Reference Palm Z22 Read this First
16
Pendragon 5.1 Software

• Pendragon 5.1 software is installed on the Palm
  Pilot and the laptop.
• Pendragon on the Palm allows you to fill out the
  survey in the field
• Pendragon Forms Manager on the laptop, stores the
  form, collects and backs-up the data

Pendragon Forms Manager on the Laptop
Pendragon Forms 5.1 for the Palm
17
How to Launch the Survey
CRM 2009
1. Tap Forms 5.1.
2. Tap FCAT 2009 and then tap New.
18
Entering the Interview Number
(overall)
Tap the number of the interview overall.
19
Review List of All Surveys Collected
Tap any survey to edit its data or continue the
survey by then tapping on survey questions.
All of the survey data you have entered for an
interview, in a scrollable list format.
Tap Review to see a list of surveys you have
collected.
20
What Is Syncing?

• Syncing is moving information between your Palm
  and the laptop. After a sync, both the Palm and
  the laptop will contain the same data

21
Title Goes Right Here
22
TRAJECTORY OF TECHNOLOGY IN MICROFINANCE
PDA
ATM
Point of Sale (POS)
Cell Phone
Technology focus shifting from data capture (for
MFIs) to more convenient transactions (for
clients)
1999
2007
23
WHY DID HANDHELD DEVICES HAVE LIMITED SUCCESS
8-10 YEARS AGO. AND HAS ANYTHING CHANGED?

• NOW
• Much higher level of MIS for many MFIs
• Some relevant technology (smartphones) already
  has high penetration in emerging markets

• BACK THEN.
• Fewer stable Management Information Systems
  (MIS)
• The technology unknown and expensive in
  developing countries

24
EXAMPLE 1 ASIRVAD IN INDIA

• WHO THEY ARE
• 50,000 members, 19 branches
• Mainly group lending
• WHAT THEY DID
• Migrated all operations to cell phones
• Eliminated computers from branches
• Put java applications in the smart phones of
  their field operations personnel (branch manager,
  loan officers) to allow them to interact with
  system
• Re-engineered operating procedures
• RESULTS
• Came up with a simple, scalable way to spread
  geographically
• Field Operations productivity rose 60 - from 500
  to 850 members per loan officer
• Grew from few thousands to 50,000 in two years

25
EXAMPLE 2 KUYASA FUND, SOUTH AFRICA

• WHO THEY ARE
• Housing microfinance institution
• Have made gt7,000 loans
• WHAT THEY DID
• Developed integrated electronic workflow that
  allows seamless link between mobile PDAs,
  accounting platform and MIS
• Mobile technology allows for loan origination in
  the field including preliminary client scoring
  (through real-time access to credit bureaus)
• Also used to receive payments (with receipts) and
  issue statements
• RESULTS
• Reduced loan origination time from 2 weeks to 3
  days
• Introduced higher level of consistency in
  decisions

26
EXAMPLE 3 JAMII BORA IN KENYA

• WHO THEY ARE
• MFI with about 200,000 members
• 72 branches and 142 outlets
• WHAT THEY DID
• Introduced handheld POS devices with which staff
  can connect clients to service at branch
  locations and in the field
• Connected MIS via GPRS over local mobile networks
• POS are equipped with biometric reader
• Re-engineered processes so nearly all
  transactions are completed via POS
• RESULTS
• Fast data processing
• Mgmt can see end of day cash position for all
  outlets
• Allowed back office to keep up with rapid
  expansion

27
HANDHELD DEVICES HAVE A MYRIAD OF POTENTIAL USES
Faster and more accurate credit approval
Reduce Fraud
Reduce Manual Error
Enforce Compliance
Reduced volume of paper
Increase productivity
Cash in/out and transfers apart from any bank
branch
Provide immediate access to info in the field
Allow field transactions to be processed real-time
Print Receipts
28
ISSUES TO CONSIDER SPECIFICALLY WHAT IS THE
BUSINESS BENEFIT?

• THE BUSINESS MUST DRIVE THE TECHNOLOGY SOLUTION,
  NOT THE OTHER WAY AROUND!!!
• Every institution is different and the impact of
  introducing handheld technology for data capture
  will differ depending on methodology and
  processes
• First, examine business processes and
  inefficiencies and define a technology solution
  to solve your specific problem.
• Some sobering data points
• SKS pilot of handheld technology in group
  meetings Time saved (5-10 min) was not worth
  cost of technology
• USAID Survey of MFIs found 80 thought PDA was
  below expectations on improved revenue and
  services and 60 below on improved productivity
  and data quality

29
ISSUES TO CONSIDER STABLE MIS PLATFORM AND
GOOD IMPLEMENTATION

• Handheld technology can make a good MIS run
  better, it will not fix or replace a weak MIS
• A well-functioning MIS is essential for
  effective results
• High quality technical support is also required
• Most technology implementations take longer than
  expected
• Dont cut corners on pilot testing and extensive
  staff training
• Consider issues like responsible parties for
  lost/damaged devices before the events occur

30
CLAUDIA MCKAY cmckay_at_worldbank.org
Advancing financial access for the worlds poor
www.cgap.org www.microfinancegateway.org
31
Better, cheaper, faster UMPC-powered data
collection

• Arthur Shaw, IRIS Center

32
A tale of two surveys, a tale of two technologies

• Liberia

• Sénégal

33
Liberia

• Survey process
• Data collection
• What asking questions, recording answers
• Who interviewers
• Data entry
• What transferring answers from paper to
  computer
• Who data entry clerks
• Data cleaning
• What correcting data entry problems, correcting
  data quality problems
• Who data analysts

34
Liberia

• Survey quality
• Data collection
• Sometimes bad information
• Sometimes no information
• Data entry
• Sometimes good information badly entered
• Sometimes no information entered
• Data cleaning
• Sometimes bad assumptions
• Sometimes bad decisions

35
Liberia

• Survey costs
• Data collection
• Staff interviewers, field supervisors
• Goods paper, printing
• Data entry
• Staff manual editors, programmer, data entry
  clerks, data entry supervisor(s)
• Goods software, computers
• Data cleaning
• Staff data analyst
• Goods software, computers

36
Liberia

• Survey timeline
• Data collection
• Training interviewers weeks
• Printing questionnaires days
• Interviewing respondents weeks
• Data entry
• Developing data entry system days
• Training data entry clerks days
• Editing questionnaires before entry days
• Entering data usually twice weeks
• Data cleaning
• Finding data entry problems days
• Correcting them with data entry team days,
  weeks
• Handling outliers, missing values, inconsistent
  answers weeks

37
Sénégal

• Survey process
• Data collection
• What ALWAYS asking THE RIGHT questions,
  recording answers IN A DIGITAL FORMAT
• Who interviewers
• Data entry
• What transferring answers from paper to
  computer
• Who data entry clerks
• Data cleaning
• What correcting data entry problems, correcting
  FEWER AND LESS DIFFICULT data quality problems
• Who data analysts

38
Sénégal

• Survey costs
• Data collection
• Staff interviewers, field supervisors
• Goods paper, printing
• Data entry
• Staff manual editors, programmer, data entry
  clerks, data entry supervisor(s)
• Goods USER-FRIENDLY software, MOBILE computers
• Data cleaning
• Staff data analyst (MUCH LESS TIME)
• Goods software, computers (MUCH LESS USE)

39
Sénégal

• Survey timeline
• Data collection
• Training interviewers weeks
• Printing questionnaires days
• Interviewing respondents weeks
• Data entry
• Developing data entry system days
• Training data entry clerks days
• Editing questionnaires before entry days
• Entering data usually twice weeks
• Data cleaning
• Finding data entry problems days
• Correcting them with data entry team days,
  weeks
• Handling outliers, missing values, inconsistent
  answers weeks DAYS

40
Promise of UMPCs

• Quality
• Accuracy - data entry occurs at data collection
• Completeness software assists in answering all
  pertinent questions
• Consistency software as super supervisor, but
  during the interview
• Cost
• Investment devices cost the equivalent of paper
  and processing of two surveys
• Savings on printing, on survey staff labor, on
  analyst time
• Time
• Data delivery updates throughout the survey,
  final data within days of last interview
• Project cycle compressed substantially,
  particularly at the end

41
Problems of UMPC

• Expertise
• Hardware purchase array of options, often
  difficult to choose
• Data entry application more complex programming
• Trouble-shooting field staff describing
  problem, home staff addressing problem
• Capacity building intuitive to use, but may
  require additional training
• Cost
• Amortization enough data collection to warrant
  investment?

42
Questions?