§ Hiring

Hiring a data-quality analyst.

Data-quality analyst is the role three teams need before any of them realizes it. Symptom: three teams are pulling the “same” number from the “same” database and producing three different answers in the executive dashboard. You don’t need more data. You need someone to own the definition.

The actual scope

• Data-quality dimensions: completeness (is the field populated?), accuracy (is it right?), consistency (does it match across systems?), timeliness (is it current?), uniqueness (no duplicates), validity (conforms to format rules).
• Owns the definition: “what counts as an active customer?” “what counts as a closed deal?” Writes the SQL, publishes the rule, defends it in meetings.
• Scheduled quality scans: weekly or nightly runs that measure each dimension against each table and produce a single dashboard.
• Remediation plans: when drift is detected, the analyst owns the investigation — where did it come from? Upstream system change? New integration? Bulk import? — and the fix.
• Cross-system reconciliation: when the CRM, the ERP, and the finance system disagree about revenue, the analyst produces the variance report and drives the conversation.

Pay

Indeed and Glassdoor national aggregates (2025) place data-quality analysts at $62,000–$98,000, median around $78,000. Higher in regulated industries — financial services ($85,000–$120,000), healthcare ($70,000–$105,000) — because data quality feeds compliance reporting and the cost of being wrong is a fine or a restatement.

The role is on a hot career path. A data-quality analyst with two or three years of experience and working SQL + Python can pivot to data engineer, analytics engineer, or data product manager at a meaningful pay bump.

Interview signals worth watching

• Can write SQL on a whiteboard without flinching. Data-quality analysts who can’t read the production database are not useful analysts.
• Has an opinion on which data-quality dimension is hardest to measure. Right answer is usually accuracy, because it requires a source of truth and accurate is always relative to something.
• Can describe a quality incident they investigated end-to-end — detection, root cause, remediation, prevention. Not just “we found dirty data and cleaned it up.”
• Comfortable pushing back on leadership. If a VP asks for a number that the data can’t support, the analyst says so. Signal: the candidate has at least one story about disagreeing with a stakeholder and winning.

Signs your org actually needs one

• Three or more teams make decisions from the same database and their numbers keep disagreeing.
• Executive dashboard numbers need a paragraph of footnote to explain.
• A regulator or auditor has already asked a question you couldn’t immediately answer.
• Migrations or integration changes happen frequently and nobody owns catching downstream quality regressions.

See also: Airtable validation, data-ops overview.