Making data quality measurable

Financial institutions have large amounts of data at their disposal that should be utilised optimally for the benefit of companies and their customers. It is therefore crucial not only to know what data is available and where it is located, but also to be able to make statements about the quality of the data. Top management in particular is dependent on this in order to be able to make well-founded decisions based on this information.

But how can data quality be measured and presented in a way that is simple and comprehensible for everyone? Our experts addressed this question as part of a project for Pfandbriefbank (pbb).

Comprehensible and reliable statements on data quality enable management to draw conclusions for well-founded decisions.

Florian Woy

Senior Manager, Financial Services

KPMG AG Wirtschaftsprüfungsgesellschaft

Data Quality Score

The aim was to develop a solution that would enable top management to gain an overview of the quality of the data for the entire bank. To this end, the "Data Quality Reporting" project was set up together with the customer. The aim was to make data quality measurable, reduce complexity and give the right impetus to the final decision-makers.

Our approach to making something as abstract as data quality visualisable was inspired by the food industry: The Nutri-Score printed on many packages gives consumers a quick and easy-to-understand impression of the nutritional values of individual food ingredients and can be a decision-making criterion when choosing a product.

Why not use something similar in the form of a data quality (DQ) score for data?

Individually calibratable approach

A scoring approach with model parameters was chosen to approach a quality characterisation of the status of the data. A quality profile for control-relevant key figures was determined from numerous sources, based on various DQ assurance measures along the data flows. The calculation and aggregation logic makes it possible to assign a data quality score at both key figure and report level. The label is intended to provide an easy-to-understand but not simplistic assessment of the quality of data.

Dynamic data quality report

Both top management and all other employees of Pfandbriefbank can view the assessments of the status of data quality in a bank-wide DQ report (dashboard), which allows for individual presentations and analyses - depending on the perspective required.

We have set up a DQ reporting system that is tailored to the target group and that everyone can use according to their needs. As it is visually structured and easy to understand, acceptance within the company is very high.

Caroline Grimm

Data Governance Officer

Deutsche Pfandbriefbank AG

You can find more information about the project for Pfandbriefbank in the following video: