Why Data Quality Matters: A Practical Guide to Governance and Seven‑Dimensional Evaluation

01 Data Governance Problem Scenarios

Business leaders often rely on dashboards to monitor KPIs. When a core report shows blank data or abnormal spikes, the cause is usually delayed upstream data or calculation errors, leading to loss of trust in the report.

02 Importance of Data Quality

High‑quality data enables accurate insight and precise execution; poor data incurs high costs, compliance risks, and bad decisions.

No department owns data‑quality issues

Requires cross‑functional collaboration

Organizations must recognize data quality as a critical issue

Needs clear data‑quality standards

Requires financial and human investment

Considered labor‑intensive

ROI is hard to quantify

Despite challenges, three reasons make data quality essential: cost, compliance, and decision‑making.

Reason 1: Cost

Poor data quality is a major cause of IT project failure and customer churn.

Reason 2: Compliance

Low‑quality data can cause legal and reputational risks, such as inaccurate credit risk assessment, incomplete credit records, and regulatory violations.

Reason 3: Decision Making

Accurate data provides timely information for product and service management; bad data leads to wrong insights and costly decisions.

03 Common Data Quality Issues

Data delay causing untimely business insights.

Data errors making results untrustworthy.

Slow data recovery and lengthy root‑cause analysis.

These issues often propagate downstream, affecting applications.

04 Root Causes of Data Quality Problems

Data platform problems: instability, insufficient queue resources causing job delays or errors.

Data development problems: poorly performing scripts or flawed logic leading to delays or incorrect calculations.

Upstream system anomalies: source system failures causing late data arrival.

05 Data Quality Governance

Effective governance requires early detection, rapid response, and quick recovery to prevent business impact.

We use a data‑quality monitoring platform to monitor Hive tables at both table and field levels.

(1) Configure Monitoring Rules – mandatory checks for high‑value jobs (e.g., primary‑key uniqueness, non‑null validation) and optional business‑specific rules (e.g., month‑over‑month checks). The platform provides ~17 field‑level and 5 table‑level rules and supports custom SQL rules.

(2) Monitoring Alerts – when a rule flags an anomaly, responsible owners are notified via phone, email, or SMS and must resolve and close the alert.

(3) End‑to‑End Data Monitoring – for high‑value jobs, data lineage is used to attach monitoring to upstream jobs, achieving full‑link quality oversight.

06 Data Quality Evaluation System

After implementing governance measures, we assess effectiveness with a seven‑dimensional model:

Data Integrity – completeness and lack of missing items (average of table and field integrity).

Monitoring Coverage – proportion of high‑value jobs under monitoring.

Alert Responsiveness – ratio of handled alerts to total alerts.

Job Accuracy – 1 – (alerted jobs / total monitored jobs).

Job Stability – 1 – (error jobs / total jobs).

Job Timeliness – 1 – (delayed high‑value jobs / total high‑value jobs).

Job Performance Score – 1 – (critical jobs / total jobs).

The model aggregates these dimensions at the database level to clarify responsibility and enable targeted remediation.

We also provide a multi‑dimensional quality report showing overall scores, trends, and drill‑down capabilities to pinpoint low‑scoring databases.

Detailed analysis includes top‑ranked low‑quality databases, encouraging owners to optimize.

Finally, a mind‑map visualizes the data‑quality governance workflow.

Data quality governance is an ongoing effort; with clear goals, responsibilities, and platform tools, organizations can build a robust data foundation that unlocks business value.