Data Governance and SSOT: Master Data Management Guide

Most companies have a “customer list” maintained in three different places: in the CRM, in the accounting software, and in the sales team’s Excel spreadsheet. Each contains different phone numbers, different discount rates, different addresses. Which data is correct? Nobody knows. This chaos is a classic symptom of a lack of data governance.

In this guide, you will find answers to questions such as: What is data governance? How do you establish a SSOT (Single Source of Truth)? What are the dimensions of data quality? How is master data managed? How is data ownership defined?

What is Data Governance?

Data governance defines who manages the organization’s data and how.

Data governance is the set of policies and processes that define how data in an organization is created, stored, shared, and protected, and by whom, within what rules.

Data governance is not solely an IT responsibility. Business units, legal, finance, operations, and senior management share joint responsibility. This is because data accuracy, security, and usage impact the strategic decisions of the entire organization.

The 4 Pillars of Data Governance

1. Data Ownership
A responsible person or unit is designated for each data set. For example, customer data is the responsibility of sales, and product data is the responsibility of product management.

2. Data Standards
Rules are established for how data should be named, formatted, and entered. For example: phone numbers must be entered in the format (0532-XXX-XX-XX).

3. Data Quality
The accuracy, completeness, consistency, and timeliness of data are continuously measured and improved.

4. Data Security
Rules are defined regarding which users can access which data, which data can be shared externally, and how personal data is protected.

The SSOT (Single Source of Truth) Concept

SSOT: A single authoritative source for every piece of data

SSOT (Single Source of Truth) is the principle that there should be a single authoritative source for any given piece of data within an organization. Other systems reference this source, do not keep a copy, or synchronize their copies if they do.

SSOT Principles

1. Single source for each data set
Customer data → CRM
Product data → PLM / ERP
Stock quantity → WMS / ERP
Price list → ERP
Supplier data → SRM / ERP

2. Other systems reference the source
An e-commerce site retrieves the customer address from the CRM via API; it does not store it in its own database. This ensures all systems are updated simultaneously when a customer’s address changes.

3. Synchronization is mandatory if copies exist
In some cases, data copies are kept for performance or offline work needs. In such situations, a real-time or scheduled synchronization mechanism is established.

Benefits of SSOT

• Eliminates data inconsistency: The sales and finance teams see the same discount rate.
• Increases trust: Reports are fed from a single source, making results indisputable.
• Reduces data update costs: Updates are made from a single location.
• Simplifies auditing: Data history is tracked in one place.
• Speeds up decision-making: Time is not wasted asking “Which data is correct?”

The Cost of Data Inconsistency

Data inconsistency incurs operational and strategic costs.

In companies lacking data governance, data inconsistency directly creates costs in daily operations and strategic decisions:

Operational Costs

Duplicate record cleanup: The same customer is registered in 3 different systems with 8 different variations. Manual cleanup takes hours.

Data validation time: The sales team requests discount approval from the finance team. Finance doesn’t trust the discount in the CRM and checks their own Excel. Extra communication, extra time for every transaction.

Incorrect shipments: A customer’s address was updated in the e-commerce system but not synchronized with the logistics software. The product is shipped to the wrong address. Return costs, customer dissatisfaction.

Report preparation time: Preparing the monthly sales report takes 2 days because data needs to be collected from 4 different sources, matched, and inconsistencies resolved.

Strategic Costs

Inaccurate forecasts: Supply planning is done incorrectly due to inconsistent stock quantities. This leads to either excess inventory costs or lost sales due to stockouts.

Loss of trust in reports: Senior management is unsure about the accuracy of reports. The question “Where does this number come from?” is asked in every presentation. The decision-making process slows down.

Deteriorated customer experience: A customer calls the call center and is directed to their old address. They call again, complain. Brand value is damaged.

Compliance risks: During a GDPR audit, it’s unknown which system contains which customer data. If a customer requests data deletion, it needs to be deleted from 5 different systems, but it might be overlooked.

Master Data Management

Master data defines the core assets of an organization.

Master Data refers to data that defines an organization’s core assets and changes infrequently. Transactional data, on the other hand, refers to the frequent operations performed on these assets.

Master Data Categories

1. Customer
Customer name, tax ID, address, contact information, payment terms, price group, sales representative.

2. Product
Product code, name, category, unit, cost, price, supplier, technical specifications, description.

3. Supplier
Supplier name, tax ID, payment terms, delivery times, quality rating, risk level.

4. Location
Warehouse, branch, facility information. Address, capacity, authorized person, region, active/inactive status.

5. Employee
Employee code, name, department, title, permissions, cost center, manager.

Principles of Master Data Management

Centralized definition: Master data is defined in a single system (usually the ERP or an MDM solution). Other systems are fed from this source.

Approval process: An approval process is followed before defining new customers, products, or suppliers. For example, finance approval is required to add a new customer.

Standard format: A format is determined for each field. Phone: (05XX) XXX-XX-XX, Tax ID: 10-digit numerical.

Regular cleanup: Master data cleanup is performed at least once a year. Duplicate or unused records are deleted or deactivated.

Version control: Master data changes are logged. It is tracked who changed what and when. If necessary, it can be reverted to a previous version.

Dimensions of Data Quality: Accuracy, Completeness, Consistency, Timeliness

Data quality is measured and tracked across 4 dimensions.

Data quality is more comprehensive than just asking “Is the data correct?”. It is evaluated across 4 dimensions:

1. Accuracy

Definition: How well the data aligns with reality.

Example: Is the customer’s phone number, 0532-XXX-XX-XX, recorded the same way in the system? Or is it an old number?

Measurement: 100 customer phone numbers are randomly dialed. The number of correct and incorrect ones is checked. Accuracy rate = number of correct records / total number of records.

Target: 95%+ accuracy rate.

2. Completeness

Definition: The extent to which mandatory fields are filled.

Example: Tax ID, address, and phone number are mandatory for customer records. Out of 1000 customers, 850 have a tax ID, and 150 do not. Completeness rate = 85%.

Measurement: The completeness of mandatory fields is queried automatically. Measured separately for each data set.

Target: 100% completeness for critical fields, 90%+ for others.

3. Consistency

Definition: The same piece of data having the same value across different systems.

Example: Customer XYZ’s discount rate is 20% in CRM, 15% in ERP, and 18% on the e-commerce site. There is inconsistency.

Measurement: The same data set is retrieved from different systems and compared. Consistency rate = number of matching records / total number of records.

Target: 98%+ consistency rate. With SSOT implementation, inconsistency approaches 100%.

4. Timeliness

Definition: How up-to-date the data is, and the delay time.

Example: A customer’s address changed. It was updated in the CRM within 5 minutes, but updated in the logistics system after 2 days. There is a 47-hour delay in the logistics system.

Measurement: The difference between the time of change and the time of synchronization is measured. Average delay time is calculated.

Target: Real-time synchronization for critical data, <1 hour delay for others.

Data Ownership and Responsibility Matrix

The most critical element of data governance is the clear definition of ownership for each data set. Data ownership resides with the relevant business unit, not IT. IT only provides the technical infrastructure.

Data Ownership Example Matrix

Roles and Responsibilities

Data Owner: Makes decisions regarding data quality, standards, and access rights. Defines business rules. Example: “Customer tax ID is mandatory.”.

Data Steward: Performs daily data quality checks. Identifies and corrects errors, or ensures they are corrected. Provides data entry training.

Technical Owner (IT): Responsible for databases, integrations, backups, security, and performance. Technically implements business rules but does not define them.

Real-World Case Study: Multi-Channel Retail Company

7 Data Governance Mistakes

1. Viewing Data Governance as an IT Project

Data governance is not an IT responsibility but a business unit responsibility. IT only provides the infrastructure. Business units define the business rules.

2. Ignoring SSOT

The approach of “Each system keeps its own data, we’ll match them when needed.” Result: Constant inconsistency, loss of trust, manual matching costs.

3. Not Measuring Data Quality

The assumption that “data is generally correct.” Accuracy, completeness, consistency, and timeliness are never measured. Problems are only noticed when they become significant.

4. Not Defining Data Ownership

The approach of “Everyone can modify any data.” Result: Ambiguity of responsibility, no quality control, no one to hold accountable for errors.

5. Neglecting Master Data

Customer, product, and supplier data are not managed centrally. Each department maintains its own list. Synchronization is impossible due to lack of integration.

6. Not Enforcing Data Entry Rules

Phone numbers are entered as “0532…” in one place, “+90 532…” in another, and “532…” elsewhere. There is no format standard, leading to low data quality.

7. Not Performing Regular Cleanup

Duplicate records, unused records, and old data remain in the system for years. Data pollution increases, and query performance degrades.

Proper data governance processes prevent mistakes.

Data Quality KPI Table

Key metrics used to measure the success of data governance:

Data Governance Checklist

Frequently Asked Questions (FAQ)