Digital Lending Analytics for Credit Risk Managers Training Course

Course overview

Digital lenders make credit decisions from application data, transaction histories, device signals, bureau files and behavioural events—often within minutes. For credit risk managers, the challenge is not simply building a more predictive score: it is setting risk appetite, proving model performance, controlling approval and pricing strategies, detecting deterioration early, and explaining decisions to model-risk, compliance and executive stakeholders. Weak monitoring can conceal adverse selection, rising first-payment default, concentration risk or unfair outcomes until losses have already accumulated.

This course develops the analytical methods used to manage digital consumer and SME credit portfolios across the lending lifecycle. Participants work with cohort and vintage analysis, scorecard and machine-learning performance measures, probability of default calibration, affordability indicators, expected credit loss concepts, champion-challenger testing, rejection inference, collections segmentation and early-warning dashboards. They learn to translate analytics into practical decisions on cut-offs, loan limits, pricing, verification rules, manual-review queues and collections interventions.

Instructor-led workshops use a realistic digital lending portfolio and structured datasets in SQL, Python and Power BI. Participants interrogate data, challenge a deteriorating model, design a test strategy and present recommendations to a mock credit risk committee. Each participant leaves with a Digital Lending Risk Analytics Playbook: a documented portfolio-monitoring pack containing KPI definitions, a vintage dashboard specification, model-monitoring thresholds, a decision-policy test plan and a 90-day implementation roadmap.

The programme is designed for experienced credit professionals moving beyond traditional bureau-led underwriting, as well as analytics leaders who need to govern automated lending decisions with commercial discipline and defensible controls.

Course objectives

By the end of this course, participants will be able to:

• Construct cohort, vintage and roll-rate analyses to isolate origination-period and policy-driven credit deterioration
• Evaluate scorecard and machine-learning models using AUC, KS, Gini, calibration curves, PSI and characteristic stability measures
• Set risk-based approval cut-offs, loan limits and pricing bands using expected-loss and risk-appetite constraints
• Design a champion-challenger experiment with hypotheses, sample rules, guardrails and success metrics
• Build SQL queries that reconcile application, decision, repayment and collections events into a lending performance dataset
• Create a Power BI portfolio dashboard for funnel conversion, first-payment default, delinquency migration and concentration risk
• Apply IFRS 9 expected credit loss concepts to segment exposures, staging triggers and forward-looking risk overlays
• Produce a model-monitoring and decision-policy governance pack for credit risk committee review

Benefits of attending

Target competencies

Who should attend

• Credit Risk Managers — who own lending policy, portfolio performance and risk appetite outcomes
• Digital Lending Managers — who need to balance approval growth, customer experience and loss performance
• Credit Risk Analysts — who monitor scorecards, vintages, delinquency and underwriting-rule effectiveness
• Head of Credit or Consumer Risk Leaders — who approve strategy changes and need defensible management information
• Model Risk Managers — who challenge model performance, drift controls and automated-decision governance
• Collections and Recoveries Managers — who use behavioural segmentation to prioritise early intervention

Requirements and prerequisites

Participants should have practical experience in consumer, SME or retail credit risk, including familiarity with application approval, delinquency measures, credit bureau data, scorecards or portfolio reporting. They should be comfortable interpreting percentages, distributions, correlation, basic probability and Excel tables, and should understand terms such as PD, default, arrears, vintage and credit policy. Prior SQL, Python or Power BI experience is useful but not essential; guided templates are provided. This is not a coding bootcamp and does not require advanced machine-learning development, prior SAS programming or actuarial modelling expertise.

Training methodology

The instructor alternates focused credit-risk briefings with hands-on analysis of a simulated digital lender’s application, repayment and collections data. Participants use SQL to assemble performance views, Python notebooks to test model and policy measures, and Power BI to communicate findings. Facilitated case discussions examine approval-growth trade-offs, model drift and fairness concerns. Small groups act as a credit risk committee, challenge proposed actions and agree controls. The final workshop converts analysis into an individual 90-day implementation plan and governance-ready monitoring pack.

Course outline

Day 1: Digital lending risk architecture and portfolio data

• Digital lending lifecycle from acquisition to recovery
• Risk appetite metrics for approval, loss, concentration and capital
• Application, bureau, open-banking and behavioural data structures
• Default definitions, observation windows and performance horizons
• Data lineage across origination, servicing and collections systems
• SQL joins for application-to-repayment performance records
• Data-quality controls for missingness, duplication and timestamp errors

Workshop: Participants map a digital lender’s data lineage and write SQL queries to create a reconciled application-to-outcome analysis table.

Day 2: Underwriting models and credit decision performance

• Scorecard construction logic and machine-learning model roles
• Discrimination metrics using ROC, AUC, KS and Gini
• Calibration assessment with observed-to-expected default rates
• Population Stability Index and characteristic stability monitoring
• Reject inference methods and selection-bias limitations
• Fairness testing across protected and proxy population segments
• Manual-review queues, verification rules and automated-decision exceptions

Workshop: Participants diagnose a declining underwriting model using AUC, calibration, PSI and segment-level performance evidence.

Day 3: Portfolio monitoring, vintages and expected loss

• Cohort and vintage curves by booking month and acquisition channel
• First-payment default and early-delinquency indicator design
• Roll-rate matrices and delinquency migration analysis
• Static-pool loss curves and maturity-adjusted comparisons
• IFRS 9 PD, LGD and EAD concepts for lending portfolios
• Stage allocation triggers and forward-looking macroeconomic overlays
• Concentration analysis by geography, product, channel and risk band

Workshop: Participants build a vintage and roll-rate monitoring pack, then identify the segments driving an adverse-loss forecast.

Day 4: Decision strategy, pricing and controlled experimentation

• Risk-based cut-offs linked to expected loss and unit economics
• Loan-limit assignment and affordability-cap design
• Risk-based pricing, APR constraints and margin-at-risk analysis
• Champion-challenger design for scorecards and policy rules
• A/B test sample sizing, randomisation and operational guardrails
• Policy simulation using approval, bad-rate and profit trade-offs
• Change-control documentation and credit committee approval criteria

Workshop: Teams design a champion-challenger test for revised cut-offs and limits, producing a decision memo with risk guardrails.

Day 5: Collections analytics, governance and action planning

• Behavioural segmentation for pre-delinquency collections treatment
• Contact-strategy optimisation and cure-rate measurement
• Early-warning indicators for portfolio deterioration
• Power BI dashboard design for credit risk committee reporting
• Model-monitoring thresholds, alerts and escalation workflows
• Model risk governance, documentation and independent challenge
• Ninety-day implementation planning for analytics-led policy changes

Workshop: Participants present a Power BI dashboard specification and complete their Digital Lending Risk Analytics Playbook for a mock credit risk committee.

Tools & standards covered

Microsoft SQL Server, Python, Microsoft Power BI, IFRS 9

A typical training day

Live online deliveries follow the same structure in the East Africa Time zone, with shorter screen blocks and longer breaks.

What the fee includes

How you can take this course

Certification

Participants who complete the full five days receive the Skillset Development Certificate of Completion, stating the course title, course code, dates and delivery format — suitable for professional-development records and employer reimbursement.

Frequently asked questions