How to Turn Lots of Data into Good Data
Why Use a Data Quality Tool?
High-quality product data boosts business success, increases sales, and naturally improves customer satisfaction by reducing the number of complaints—many of which are caused by incorrect product descriptions.
Poor data quality, often due to incorrect or incomplete product information, can even harm a company’s credibility. Studies show that bad data can negatively impact a brand or manufacturer.
In many cases, complete data is the key to better search visibility for products on the web—which in turn drives higher sales figures.
Using tools to validate data helps avoid wasting resources—for example, in customer service or through manual correction of data errors. This leads to cost savings and greater process efficiency within the company.
In short, a data quality tool offers extensive monitoring capabilities, making it possible to optimize content and data maintenance processes either per project or on an ongoing basis.
Key Validation Criteria for High-Quality Data
The following validation criteria define professional data quality, making them essential—and, of course, they are included in the Data Quality Tool:
Completeness
Every attribute expected by a consuming system must be checked for completeness. The exact number of missing data points is calculated per evaluation and displayed as a percentage. Attributes still requiring input can be quickly and easily identified using a filter function.
Example: Product information for Evaluation A = 80% complete; the remaining 20% of attributes can be displayed directly.
Accuracy
Even with 100% data completeness, errors can still occur during the initial data entry. Each attribute and data type can be evaluated from different perspectives. Specialized validation logic is applied specifically for this purpose. The search results directly highlight critical cases, which can then be reviewed and corrected as quickly as possible.
Example: Incorrect entry = Weight must be > 0
Uniqueness and Consistency
Existing data records must not conflict with one another and should be stored in a consistent format across the system. This type of validation is often underestimated, yet it is just as important as the other quality factors. One scenario you definitely want to avoid is publishing identical articles that differ in content. This kind of consistency check can also be applied to other data sets.
Example: Duplicate of file variant A for product image B
Recency and Validity of Ordering Information
The tool not only performs content-related checks but also takes into account the recency and validity of ordering information. With the additional Portfolio Check, specific publication criteria can be validated based on regional assortments.
Criteria that are typically consistent across use cases include:
- Does the data record exist?
- Is the data record approved?
- Is the data record available for sale in the given region or country?
- Is there a valid price associated with the data record?
Example: Item number A = Status: available for sale
The report generated after evaluation is available as an Excel file, showing the specific results for each item number.
Incomplete Data – What Now?
Checking the data alone is not enough—it’s important to take timely action. That’s why the Data Quality Tool offers various features to support the follow-up process.
A granular permission management system allows different user groups to access the tool, with access to specific content and functions depending on their assigned role. Administrators not only see all existing evaluations, but can also approve data records for publication once all data has been corrected and validated.
With the download function available on every page, each data record and its contents can be exported as a Microsoft Excel file and passed on to the responsible team for further processing. The file includes not only the core information but also metadata on the data source—since the information typically comes from multiple systems (ERP, TecDoc, PIM, MAM…).
The Data Quality Tool as a Universal Solution for Multichannel Systems
The definition of which data records and attributes should be checked by the Data Quality Tool—and against which criteria—always depends on the sales organization, language, and specific use case.
Not all content is relevant to every use case. For example, printed catalogs require different file versions of product images than digital sales channels. This means individual requirements can be defined and validated for each evaluation scenario.