Wells Fargo · Primly Community

Wells Fargo data scientist interview (SQL + case + stats): full breakdown

de_derek · 5 replies

Wrapped a DS loop at Wells Fargo targeting their consumer analytics team. Level was mid-senior (5 YOE). Here's every section of the interview with what actually got asked.

SQL round (45 min): This was the hardest SQL I've seen outside of a pure analytics engineering role. They do window functions, partitioning, running totals, lag/lead. The specific problem was something like: given a transaction table with customerid, amount, transactiondate, and transactiontype, find customers who had a transaction decline followed by a successful transaction within 72 hours. They want to see you think through the join logic and handle edge cases like multiple declines.

They use Snowflake internally. Syntax matters. Know that DATEDIFF and DATEDIFF behave differently in Snowflake vs SQL Server.

Stats / probability round (30 min): Not a stats theory test. They asked about a real scenario: we're seeing fraud rates spike on mobile transactions. Walk me through how you'd diagnose whether this is a true trend or a sampling artifact. Classic A/B testing fundamentals, power, significance, but framed as a detective problem. They pushed on: how would you control for the fact that mobile transactions skew toward younger customers who might just have different fraud patterns anyway?

Case / business round (45 min): My case: a new credit card product has launched and early activation rates (customers who get the card and actually use it within 30 days) are below forecast. What might explain this and how would you investigate? I structured it as: funnel analysis, cohort comparison, external factors, and proposed a decision tree of what data I'd pull first. They liked that I started with diagnostic framing before jumping to hypotheses.

Behavioral (30 min): One question that surprised me: tell me about a time your analysis led to a decision that turned out to be wrong. What did you do? WF culture stuff again. Have a real answer for this.

Comp offer I received: $148k base + $20k sign-on + 8% target bonus. Charlotte, hybrid 3 days in-office. Accepted.

5 replies

analyst_ana

The fraud spike diagnostic question is a great one. Did they give you any specific metrics or was it fully open-ended? Like did they tell you the rate went from 1.2% to 2.1% or just 'fraud rates are higher on mobile'?

content_cole

Gave me rough numbers: mobile fraud rate increased from about 0.8% to 1.6% over three months. Online (desktop) stayed flat at 0.5%. That's enough to work with. The key insight they wanted was that a 2x increase in rate sounds alarming but you need to know if mobile transaction volume also changed a lot over the same period, which it did. So the absolute fraud count increase was larger than the rate suggests.

ml_mike

The activation rate case is classic credit card analytics. Worth knowing the standard industry benchmarks going in: healthy activation rates vary by product type and acquisition channel. If you're going for a financial services DS role and you haven't read a little about credit card economics, that's a gap to close before the interview.

finance_faye

$148k base + sign-on in Charlotte for a mid-senior DS role is solid. Charlotte has a lower cost of living than NYC or SF so the purchasing power is real. What level did they map you to internally?

pivot_pat

Mapped to something like Band 4 or Senior Analyst II depending on how you count it. WF has their own internal leveling that doesn't map cleanly to FAANG frameworks. The recruiter described it as 'senior individual contributor, not yet manager track but with clear promotion pathway.'