You moved to ERP, bought a new POS or set up a reporting dashboard; but the numbers do not match, the same customer is recorded three times and products appear with different codes. The problem is not the software but the data. Data quality and master data management are the discipline that keeps a business's core records correct, consistent and current. Reporting, automation, AI and integration cannot give reliable results if the data is unreliable.
What is master data?
Master data is the core records a business uses repeatedly across many processes:
- Customers and suppliers: Name, contact, tax details, addresses, terms.
- Products and services: Code, name, category, unit, price, cost, barcode, variants.
- Employees and roles.
- Locations, warehouses, branches.
- Chart of accounts and categories.
These differ from transaction data (sales, orders, movements), but the accuracy of transaction data depends on master data.
Data quality dimensions
- Accuracy: Does the record reflect reality?
- Completeness: Are required fields filled?
- Consistency: Is the same information the same across systems and records? Are spellings standard?
- Uniqueness: Are there duplicate records?
- Timeliness: Is the data current or outdated?
- Validity: Does it comply with format and rules (phone, email, tax number, date)?
The cost of low data quality
- Wrong decisions: Faulty reports and forecasts; see demand forecasting and stock planning and KPIs and management dashboards.
- Operational errors: Wrong shipment, double invoice, unpaid receivable, wrong stock; see retail stock tracking and the e-Invoice guide.
- Inefficiency: Staff time goes to fixing errors.
- Customer experience: Double emails to the same customer, wrong greeting, missed information; see the CRM guide and email marketing guide.
- Integration and automation problems: Records that do not match between systems; see API integration and AI process automation.
- AI and analytics: Models trained or fed with bad data give wrong results; see where to start with AI.
- Compliance risk: Old or wrong personal data, inability to meet deletion requests; see the KVKK compliance checklist.
- ERP project failure: A project starting with bad data carries the problems over; see why ERP projects fail.
A data quality programme step by step
1. Scope and priority
First choose the most critical master data areas (for example customer and product). Measure the business impact of data quality problems (how many wrong invoices, how many duplicate records).
2. Data inventory and ownership
Where is data held (ERP, POS, CRM, e-commerce, Excel)? Assign an owner (data steward) for each master data area; make authority to create and change records clear.
3. Measure data quality (profiling)
Establish the baseline with simple indicators such as duplicate record rate, empty field rate, number of invalid formats and inconsistent spellings.
4. Define standards
- Coding and naming rules: Product code logic, name format; see barcode and label systems.
- Required fields and formats: Phone, email, address, tax number, unit.
- Category and classification lists: Choice lists instead of free text.
- Duplicate definition: On which fields is a record considered the same?
5. Clean up
- Merge duplicates: Without losing transaction history.
- Complete missing fields and fix formats.
- Archive old or unused records.
- Match source systems: A common key for the same customer or product in different systems.
- When moving data to another system, clean first; see moving from Excel to business software.
6. Prevent at the source
If cleanup stays one-off, data degrades again. Preventive controls:
- Validation at entry (format, required field, choosing from a list),
- Duplicate checks when a new record is created,
- Authorisation and approval flow (for master data changes),
- Automatic synchronisation between systems; integration instead of manual double entry; see digital workflows,
- Tools that reduce error at entry such as document and invoice reading; see AI document processing.
7. Monitor and sustain
Measure data quality indicators regularly, report and find the source of the error to fix the process; see reporting with Power BI. Periodic data cleanup and meetings with owners.
Security, privacy and data governance
- Access: Limit master data changes to authorised roles; keep action logs; MFA.
- Personal data: Do not collect unnecessary data; retention period and deletion process; see the data protection guide.
- Backup: Back up before cleanup and bulk changes; see backup strategy.
- Supplier data: Data in external systems should be subject to the same standards; see supplier and third-party cyber risk management.
Sector examples
- Retail: Product code, barcode, variant, price and stock consistency; see choosing POS software for a store and omnichannel retail.
- Restaurants: Recipe, ingredient codes and units; see restaurant cost control.
- Hotels: Duplicate profiles in guest records, preferences and history; see hotel management system and hotel upselling.
- Accounting: Cleaning customer and product cards; see choosing accounting software.
Common mistakes
- Thinking data problems are "software bugs",
- Doing cleanup once and not preventing at the source,
- Not assigning data owners,
- Not setting standards (code, naming),
- Moving to a new system without cleaning,
- Corrupting transaction history while merging duplicates,
- Not measuring,
- Needless accumulation of personal data.
A starter plan
- Choose the 1–2 most critical master data areas (customer and product).
- Measure the current state (duplicate, missing, inconsistent rate).
- Define standards and owners.
- Back up and clean.
- Add entry checks and automatic synchronisation.
- Track quality indicators monthly.
ujans runs data cleanup, migration, integration and reporting together and builds a reliable data foundation for AI and analytics.