Data QualityThe Foundation of Reliable Procurement Processes

The Classic Quality Dimensions

Usually five to six dimensions are distinguished: completeness (all mandatory fields populated), correctness (data matches reality), currency (data is up to date), consistency (data is consistent across systems) and unambiguity (each record exists only once). For example, an item can be complete yet still incorrectly maintained if the unit

• Completeness: mandatory fields such as unit, plant and material group are populated
• Correctness: details match manufacturer specifications and test reports
• Currency: prices and delivery times are checked regularly
• Unambiguity: every item and supplier exists only once in the system

Distinction from Data Clean-up and Data Monitoring

Data clean-up refers to the subsequent correction of faulty stocks, for example through duplicate detection or populating missing fields. Data monitoring, by contrast, is the ongoing review and approval of new data. Together, both secure data quality: clean-up remediates existing stocks, monitoring prevents new errors from occurring.

Importance for Technical Procurement

In technical procurement, data is particularly complex: material numbers, drawing numbers, works standards, customs tariff numbers and certificates must match each other. A single incorrect material entry can trigger a production batch that later becomes unusable. Data quality is therefore not an IT issue but a core component of process reliability.

Missing Responsibilities in the Master Data Process

Without a clearly regulated master data process, employees from sales, planning or production create items or suppliers on their own initiative, often with incomplete information. There is no central role that reviews entries, countersigns standards and enforces mandatory fields. In practice, it has proven effective to submit requests via a form or workflow tool in which mandatory information is enforced technically.

• Clear creation request with mandatory fields
• Four-eyes review by a master data role
• Approval levels depending on data type

Free-Text Orders and Duplicates

When buyers order past the material number and catalogue, duplicates and costly incorrect purchases arise. Consistent duplicate detection at creation and regular reconciliation of existing stocks prevent the same item from being maintained under five material numbers. This also increases the significance of analyses and lowers the stock value.

System Breaks and Interfaces

Many companies maintain data in parallel in ERP, PLM and supplier portals. Without coordinated interfaces, inconsistencies arise: one price in the portal, another in the ERP. Consistency checks between systems are therefore part of ongoing data controlling, not one-off projects.

Data Analysis in Procurement as a Starting Point

The starting point is a systematic data analysis in procurement: how many materials have no unit, no material group or no current price? How many supplier master records are inactive, duplicated or without a bank account verification? Such analyses can be created with standard reports or simple queries and provide a prioritised picture of the error quantities.

• Measure error rates per record type
• Prioritise critical fields (prices, units, certificates)
• Present the results to management as a baseline

Data Clean-up and Duplicate Detection

During data clean-up, duplicates are merged, incorrect entries corrected and outdated records archived. Duplicate detection should go beyond exact matches and also check similarities in descriptions or manufacturer part numbers. Clean-ups should preferably run in stages: first master data that impedes ongoing processes, then historical data.

• Step-by-step clean-up according to process relevance
• Similarity checks across descriptions and part numbers
• Document the change history

Prevention Through Data Monitoring and Standards

Lasting data quality is created through prevention: binding naming conventions, enforced mandatory fields, periodic data monitoring and clear approval rules in the master data process. New supplier master data should be checked including identifiers such as the DUNS number, so that credit and risk analyses are based on clean foundations.

Completeness and Error Rates

The completeness rate indicates what percentage of the relevant records has all mandatory fields correctly populated. In well-managed systems, material master data typically scores above 95 per cent, while unmaintained stocks are significantly lower. In addition, the error rate per field is useful for targeted maintenance.

• Completeness rate per record type
• Error rate per critical field
• Number of open correction orders

Duplicate and Share Rates

The duplicate rate measures what percentage of material or supplier master records exists in duplicate. In addition, it is often surveyed what percentage of procurement volume is handled via clean, catalogued materials, and what share is spend under management. Both show whether processes and data structure are effective.

Currency and Process Impact

In the medium term, data quality shows in process metrics: share of automated invoice approvals in the invoice approval workflow, rate of error-free purchase orders, lead time of invoice clearing. If these values rise after a clean-up, this confirms the economic benefit of the effort.

Operational Risks in Day-to-Day Business

Incorrect planning parameters such as safety stock or replenishment lead time lead to excess stock or bottlenecks. A safety capacity can only be agreed sensibly if the underlying consumption and lead time data are correct. An ABC/XYZ classification is likewise only as reliable as the underlying movement data.

Supplier Risk and Master Data Integrity

Incomplete supplier master data complicates the assessment of supplier risk: missing certificates, outdated addresses or unclear supplier status entries prevent reliable risk analyses. Particularly critical are regulatory evidence such as the long-term supplier declaration (LLE), whose documentation only works with correctly maintained supplier and material assignments.

• Link certificates and evidence to the supplier master
• Keep status and change data up to date
• Use identifiers such as the DUNS number for unambiguous identification

Controls in the Ordering and Invoicing Process

Automated checks in the ordering process and in the invoice approval workflow act as an early warning system: if invoices frequently go into manual clarification due to price differences or missing order references, this indicates data gaps. Regular evaluations of these clarification cases provide concrete pointers as to which records need to be corrected.

AI and Data Analysis in Procurement

Modern tools for data analysis in procurement can automatically normalise descriptions, suggest duplicates and supplement missing classifications. This considerably accelerates clean-up projects but does not replace expert review. Especially for technical items, approval by qualified staff remains indispensable.

• Automatic normalisation of item descriptions
• Suggestion function for classifications
• Human approval remains mandatory

Linking End-to-End Processes with Suppliers

Data quality becomes even more important when processes are directly interlinked with suppliers, for example in dock-to-stock delivery concepts or consignment stock. Such models require precisely coordinated item and quantity data, as corrections during the material flow are hardly possible. Order process optimisation therefore almost always begins with a clean-up of master data.

Data Quality as a Building Block of the Procurement Strategy

Strategic instruments such as bottleneck analysis, the assessment of an A-supplier or single-source management presuppose that analyses are reliable. Where data is incomplete, risks are misjudged. Leading procurement organisations therefore treat data quality as a standalone sub-objective of their procurement strategy with a fixed budget and clear responsibility.