Poor customer data quality measurement can be no worse than no measurement at all
What our approach delivers:
- Provides a simple but multi-dimensional view of customer data quality
- Based on clear linkages through to client priorities and strategies
- Creates structure and focus without constraining effectiveness
- Supports visual dashboard reporting and logical drill-down
- Uses combinations of actual measures for accuracy and indexes for clarity
How our approach works:
- Customer Data Quality measures cannot be optional
Even in organisations that fully accept the criticality of high quality customer data, meaningful improvement is often limited by measures that are not fit for purpose. Data owners may, for instance, be clear on the improvements achieved in collecting dates of birth but have completely missed the fact that an increasing number of sales records are not linking to the right customers.
- Structured, focussed and relevant to each client
Our approach to building a data quality measures framework always starts with an assessment of what is important to the client rather than with a generic list of things that can be measured. We aim to create focus in our frameworks so that they are relevant in coverage and manageable in terms of their scope.
- Data quality cannot be measured in one dimension
For each type of data identified as requiring coverage in the framework we create its own measures that focus, in varying degrees, on Data Field Quality, Whole-Record Quality, Data Integrity and Duplication Levels.
- Having just “something” in a field doesn’t count
Field-level quality can never be ignored even though it rarely tells a whole story. It is also critical to recognise that simply counting ‘entries’ in a data field, irrespective of the content, is useless. Our frameworks always create a clear quality definition for each measured field. If the field is not populated to the quality level then, on our frameworks, it is not populated at all.
- Customers have ‘Records’ not ‘Fields’
The quality of whole records is almost always more insightful than the quality of individual fields. It is easily possible to have 60% population of a number of key fields but to have a large number of records with none of these key fields populated. Without whole-record quality measures this could easily be missed.
- Even perfectly populated records may be useless
Data integrity measures are often the most important of all. The integrity measures will be the most tailored for each client of all the measures. They can work within individual data types or across data types. They could identify the proportion of sales records that don’t link to a current customer or the number of customers with a status that doesn’t match the number of agreements that they have. Both of these can have major impacts on business performance measurement accuracy.
- Duplicated records impact customers too
Record duplication has a business impact but also has substantial customer impact. Customer value measurement, segmentation and loyalty recognition are only possible if all the data for a customer is linked to a single customer record.