Paytm Data Analyst Interview Questions

Cracking the Paytm Data Analyst Interview

Paytm (based in Delhi NCR) is one of India's largest FinTech ecosystems. As a Data Analyst here, you are dealing with millions of micro-transactions daily. The focus is heavily on transaction success rates, fraud detection, and cross-selling financial products (insurance, mutual funds) to the existing user base.

The Interview DNA

Paytm interviews are fast-paced and heavily index on root-cause analysis (RCA). They want analysts who can quickly identify why a payment failed and whose fault it was (the bank, the network, or the app).

Round 1: SQL and Data Extraction (Live Coding)

FinTech runs on SQL.

• Question: "Given a table of transactions with columns user_id, timestamp, amount, and status, write a query to find the users who had 3 consecutive failed transactions within 10 minutes."
• Focus: This requires advanced Window Functions (LAG/LEAD) or complex self-joins. You must write optimized code because Paytm's actual data tables are astronomically large.

Round 2: The Root Cause Analysis (RCA) Case

This is the core of the Paytm Data Analyst role.

• The Scenario: "At 8:00 PM last night, the UPI transaction success rate dropped from 98% to 92%. How do you find the cause?"
• The Approach: Be structured.
  1. Slice by Bank (Is SBI down?)
  2. Slice by Telecom Operator (Is Jio network dropping packets?)
  3. Slice by App Version (Did a new iOS update break the payment button?)
  4. Slice by Payment Type (Is P2P failing, or P2M - Person to Merchant?)

Round 3: Product Analytics & Cross-selling

Paytm makes money by upselling services.

• Question: "How would you identify which Paytm wallet users are most likely to buy Paytm Insurance?"
• Focus: Discuss predictive modeling features. Do they frequently pay hospital bills? Do they book flight tickets often (travel insurance)? Show how you would build a target profile using existing data.

Round 4: Risk and Fraud Detection

• Question: "How would you define a logic to flag a potentially fraudulent merchant account?"
• Focus: Look for anomalies. High transaction volumes at 3 AM, hundreds of payments from the exact same IP address, or a sudden spike in chargeback requests.

Round 5: Behavioral & Fit

• "Tell me about a time you had to push back on a business request because the data quality was too poor to make a decision."
• "How do you handle the pressure of real-time monitoring when a system goes down?"

Pro Tip: Read up on how the NPCI (National Payments Corporation of India) routes UPI transactions. Understanding the architecture of a payment (Sender Bank -> PSP -> NPCI -> Receiver Bank) will make your RCA answers significantly more realistic and impressive.