Data quality

What it means

Data quality expresses the extent to which data are accurate, complete, current, consistent and free of duplicates. Data are of good quality when they reliably serve their purpose, whether invoicing, reporting, marketing or management decisions.

The quality of master data about customers, suppliers, products and employees matters most. A wrong address or VAT number leads to an incorrect invoice, a duplicate customer to split receivables, and a wrong product unit to errors in stock records.

Data quality is assessed along dimensions such as accuracy, completeness, timeliness, uniqueness and conformity to rules. The wider framework of data governance also covers who is responsible for which data and how errors are corrected so that they do not recur.

How it works in practice

The cheapest approach is to prevent errors at entry. Forms in an ERP or CRM can validate the format of company and VAT numbers, look them up in a register and search for existing records to avoid duplicates. Mandatory fields and drop-down lists reduce typos.

Regular data cleansing involves finding duplicates, filling in missing information and archiving inactive records. Before a major change such as a data migration to a new system, cleansing is essential; otherwise old errors are carried into the new environment.

For accounting, good data are the basis of reliable reports. If customers, cost centres and jobs are recorded inconsistently, management reports show wrong margins and profitability. Simple naming rules and a responsible data owner save hours at period end.

What to watch out for

Data quality also has a legal side. GDPR requires personal data to be accurate and kept up to date, and data subjects have a right to rectification. Outdated contacts in marketing databases can lead to breaches, for example by contacting someone who has withdrawn consent.

Data quality problems often surface late, during an inspection or audit, or when automation or artificial intelligence is introduced and multiplies the errors. It is worth tracking a few simple indicators, such as the share of duplicate or incomplete records, and reviewing them regularly.