Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Churn Analysis for Microfinance

Churn Analysis for Microfinance

An ML-based web app to predict whether a customer currently paying off a loan will stay with the company to get another loan or not

Shanelle Recheta

March 22, 2021
Tweet

More Decks by Shanelle Recheta

Other Decks in Business

Transcript

  1. F
    CHURN
    ANALYSIS
    01
    for a Microfinance Institution
    TEAM FAST AND FURR-IOUS

    View full-size slide

  2. F
    02
    Alleviate
    Empower
    Grow
    Providing financial and non-financial services
    to improve the economic well-being of the
    poor.
    TEAM FAST AND FURR-IOUS

    View full-size slide

  3. 03
    Why are clients defaulting?
    Why are they dropping out?
    "To serve them,
    we need to retain them."
    F TEAM FAST AND FURR-IOUS

    View full-size slide

  4. F
    04
    Improved Products
    and Services
    Reach target loan
    portfolio by 2023.
    More Happy Clients,
    More Loyal Clients
    Decrease in client dropouts
    Growth and More
    Lives Changed
    Reach more clients,
    touch millions of lives
    by 2023.
    TEAM FAST AND FURR-IOUS
    Value Proposition

    View full-size slide

  5. 05
    1 2 3 4 5
    50
    40
    30
    20
    10
    0
    F
    Tenure in Years
    ● Business Intelligence
    ● Descriptive Analytics
    ● Inferential Analysis
    To study the key factors
    related to dropout cases
    from the identified products
    of the company using:

    View full-size slide

  6. 06
    Underlying Magic
    Extract
    Retrieve data from the
    database using
    Microsoft SQL.
    Explore
    Gather initial intel on the
    structure of the data
    using Excel,Tableau and
    and Python
    Visualize
    Extract insights and tell
    the story of the data
    using Tableau
    F
    Clean
    Data preparation
    through the use of
    Python
    Communicate
    Present findings
    to the stakeholders
    using Tableau
    and Powerpoint
    Deploy
    Turnover of the final data
    set, notebooks, tableau
    files, and presentations
    to our client
    Delivered minimum viable products while continuously getting feedback from client on how to further improve the project.
    Ways of working: Agile Methodology
    SQL

    View full-size slide

  7. 07 Domain
    Knowledge
    Data
    Issues & Validation
    F
    Challenges Encountered
    Data Wrangling
    Techniques
    Short
    Timeline
    SQL, Tableau,
    Python
    Tools

    View full-size slide

  8. 09
    F
    Data issues uncovered
    helped the client realize the
    need for better structure of
    data and design of the system
    light-bulb ideas
    Actionable Insights
    and gave them
    on how to improve their system
    application control and data collection

    View full-size slide

  9. 08
    F
    Actionable Insights
    42%
    DROPOUT CLIENTS
    7M
    UNCOLLECTED PAYMENTS
    2 out of 5 clients dropout
    Client’s first loan and year
    are critical stages in retaining
    them.

    View full-size slide

  10. 10
    F
    Actionable Insights
    New clients
    are more sensitive in their experience hence,
    higher tendency to drop out.

    View full-size slide

  11. 11
    F
    It’s not just about the product.
    It’s also about customer experience.
    Actionable Insights

    View full-size slide

  12. 11
    Design Referral System
    Enhance Data Collection
    ● Train staff as “brand ambassadors”
    ● Optimize process for convenience
    Conduct Further Studies
    ● Market segmentation
    ● Social media analytics
    F Recommendations
    Improve Customer Experience
    ● Collect more customer data
    ● Regular Customer Satisfaction Surveys
    ● Data Validation Controls
    ● Encourage customers to refer
    ● Make it easy and rewarding

    View full-size slide

  13. 12
    Team Fast
    and Furr-ious
    Grace
    Database Master, Query Specialist
    Queenie
    Pythonista, Data Wrangling Expert
    Angel
    Pythonista, Data Wrangling Expert
    Airees
    Tita of Tableau, Visualization Specialist
    F

    View full-size slide