A regular dose of D2C-centric resources & tools for Growing Brands, Startups & Entrepreneurs.

Managing suspicious customers is essential to reduce potential losses, but how do you flag them?

This issue explores technical strategies for identifying and monitoring suspicious behaviour, aka SUStomer Management — providing actionable insights for a loss-free business.

1. Cross-Brand Behavioral Analytics

Utilising data across multiple brands (500+ Indian D2Cs), advanced behavioural analytics can identify complex patterns indicative of fraud:

• Cross-Brand Purchase Analysis: Detects customers making similar high-risk purchases across multiple brands.

• Anomalous Shopping Behaviour: Identify deviations from typical shopping patterns using data from a broad customer base.

• Time-Based Purchase Patterns: Analyse the timing of purchases to detect unusual patterns, such as purchases made during odd hours consistently.

2. COD Order Cancellation Monitoring

Monitoring COD order cancellations across brands to detect suspicious activities:

• Cancellation Frequency: Flag customers with a pattern of frequent cancellation of COD orders.

• Reason Analysis: Analyse reasons for COD order cancellations, such as refusal at delivery or change of mind.

• COD Risk Score Calculation: Assign risk scores based on the frequency and reasons for COD order cancellations.

3. Disabling COD for High-Risk Customers

Using cross-brand data, we can implement strategies like disabling Cash on Delivery (COD) for high-risk customers:

• Risk Scoring: Assigns a risk score based on historical data from multiple brands.

• COD Disabling: Automatically disables COD for customers with high-risk scores, even when they purchase from a new brand.

• Past Data Utilisation: Leverages past fraud data to make informed decisions about current transactions.

4. Detection During Checkout

Detecting suspicious customers at the checkout stage is essential to prevent fraudulent transactions. Effective techniques include:

• Behavioural Analytics: Monitor unusual shopping patterns, such as high-value purchases from new accounts.

• Device Fingerprinting: Identify and track devices used in transactions to spot suspicious activities.

• IP Address Analysis: Detects inconsistencies such as mismatched billing and shipping addresses or locations known for fraud.

• Velocity Checks: Identifying rapid succession of purchases that deviate from normal customer behaviour.

5. Cross-Brand Refund/Exchange Request Analysis

Utilising data from multiple brands to identify patterns of excessive refund or exchange requests:

• Refund/Exchange Frequency: Flag customers with a history of frequent refund or exchange requests across different brands.

• Product-Specific Return Analysis: Flag customers who consistently return specific types of products across multiple brands.

• Cluster Analysis: Group customers based on similar refund/exchange behaviour and identify outliers.

• Shared Refund/Exchange Blacklists: Utilise shared databases to track customers known for repeat refund abuse across brands.

6. AOV Analysis for First-Time Customers

Analysing the Average Order Value (AOV) for first-time customers across brands to detect potential trial-based fraud:

• High AOV for First-Time Purchases: Flag customers who make purchases with significantly high AOV for their first transaction, and disable COD option

• Product Diversity Analysis: Identify customers who purchase a diverse range of high-value products for their first order as it could be a Trial-Focused purchase. Meaning, there is a high probability of return.  

7. Post-Purchase Monitoring with Cross-Brand Data

Monitoring customer activities post-purchase can help detect and mitigate fraud that occurs after the transaction. Effective strategies include:

• Cross-Brand Return Patterns: Identify patterns of excessive returns across multiple brands that may indicate fraudulent behaviour.

• Shared Customer Blacklists: Utilise shared blacklists of known fraudulent customers across brands to prevent repeat offenders.

• Order Tracking and Alerts: Monitor high-value orders and send alerts for suspicious activities.

• Customer Feedback Analysis: Analyses reviews and feedback for signs of fraudulent behaviour.

• Account Activity Monitoring: Regularly check for unusual activities in customer accounts, such as multiple failed login attempts or sudden changes in order patterns.

8. Peer Group Analysis

Analyse customer behaviour in the context of peer groups to identify outliers and potential fraud:

• Cluster Analysis: Group customers with similar behaviours and identify outliers.

• Peer Group Comparison: Compare individual customer behaviour against peer group norms.

• Suspicious Activity Scoring: Assign scores to customers based on deviations from peer group behaviour.

9. Limiting COD Losses: Adaptive Measures

Implementing adaptive strategies to limit COD losses while ensuring customer satisfaction:

• High-Risk Customer COD Disabling: Automatically disable COD for customers flagged as high-risk based on cross-brand data insights.

• COD Order Verification: Implement additional verification steps, such as OTP confirmation, for COD orders from high-risk customers.

• Dynamic COD Limits: Adjust COD limits dynamically based on the risk profile of the customer and transaction.

10. Data-Driven Insights and Custom Alerts

Leverage extensive data to generate insights and custom alerts for potential fraudulent activities:

• Custom Alert Generation: Create tailored alerts based on specific fraud patterns detected across multiple brands.

• Data Visualisation Dashboards: Provide real-time visualisation of potential fraud activities and trends.

• Proactive Fraud Detection: Use predictive analytics to proactively identify potential fraud before it occurs.

That’s the end of our talk on “SUStomer Management.”...

☕ See you on the next coffee date!

How did you like our Newsletter? What topic would you like us to cover next?

Reply and let us know! 😊