Two teams report different revenue for the same day. Both queries are correct.
Whether the candidate recognises a governance problem wearing the costume of a bug, and whether they can decompose the gap before arguing about it.
The situation behind the question
Interviewers ask this because it happened to them.
Finance and growth each present a figure for last month. The queries have been reviewed and neither contains an error. The executive who received both wants to know which one is right.
A strong answer
Flags
Green flags
- Starts from semantics and consumers rather than from the SQL.
- Decomposes the gap by dimension before anyone argues, which usually ends the argument.
- Knows the owner has to be someone accountable for the number, not the engineer who can implement either.
- Adds a parity test, converting a slow social discovery into a fast automated one.
Red flags
- Diffs the two queries, picks the cleaner one, and closes the ticket. The numbers match and the condition that produced the divergence is untouched.
- Asserts the smaller figure is right because it is more conservative.
- Believes a semantic layer prevents two teams from genuinely needing two definitions.
- Does not consider that both figures might be wrong in the same direction.
Follow-ups
Where the conversation goes if the first answer holds up.
- Both figures turn out to be wrong because they share an upstream model that double counts. How would you have found that?
- The definition changes. What happens to the trend chart that spans the change?
- How do you make the shared definition the path of least resistance rather than a rule people route around?