Who owns data quality?
Whether the candidate can separate the obligation to produce correct data from the obligation to measure it, and whether they treat quality as engineering rather than as diligence.
The situation behind the question
Interviewers ask this because it happened to them.
A dashboard is wrong because an upstream field changed meaning. The data team is asked why they did not catch it. The producing team says they were not told anyone depended on that field.
A strong answer
Flags
Green flags
- Treats quality as engineering: named assertions, named owners, agreed consequences.
- Distinguishes the producer's obligation from the platform's and can defend the line.
- Assigns ownership to a team with a rota rather than to an individual.
- Knows that a muted failing test is worse than an absent one, because it reports a standard nobody enforces.
Red flags
- Says the data team owns quality, which makes the obligation unfulfillable and the failure certain.
- Proposes a policy with no mechanism that makes the supported path easier than the workaround.
- Cannot name what a producing team would actually be signing up to.
- Treats ownership as a cultural problem rather than as a required field with a gate behind it.
Follow-ups
Where the conversation goes if the first answer holds up.
- The producing team will not agree to a contract. What is your next move?
- A dataset has no owner and three consumers. How do you resolve that this week?
- What is the smallest set of obligations you would ask a producing team to accept?