Data quality – A matter of life and death?

The subject of data quality and the fast and furious world of motor racing may seem worlds apart, yet as I watched a particularly exciting motorcycle race recently, it occurred to me just how vital it is to the riders and drivers in any motorsport that the data collected and used by their engineering teams is supremely precise.

To use a scenario more relevant to most of us, how would you feel about getting on an aeroplane if you thought that the information being exchanged between the control tower, the computers and the pilots wasn’t accurate, or being interpreted incorrectly?

The quality of your customer data may not have such a potentially dramatic impact on life and limb, but it does all boil down to trust – if you don’t have confidence that your information is accurate, then you won’t trust the analysis and reports you are generating. Here are a couple of ideas you could consider to improve that trust:

Staff engagement (‘crowdsourcing’) – How this is achieved within your organisation will depend on your systems and processes, but providing a quick and easy way for staff to raise issues as they are found is a good way of improving quality over time. It also has the benefit of ensuring each staff member feels they can make a difference.

Customer input (progressive profiling) – Where certain details for existing customers are missing entirely, progressive profiling provides a method of requesting specific pieces of information from them when they next visit your site. The benefit of this method is that users are only being asked to provide new, relevant information, as their existing details are dynamically pre-populated for them on your site. For example, if you already hold email details for a particular user, on their next visit to your site you could ask for first/last name details, leaving further info (such as interests or phone number) for another day.

Another idea to consider is MasterVision DQ, the data quality module for MasterVision. Specifically developed to identify issues in your data sources, it also provides information to help you fix those issues at source, as well as measuring improvements you make to the quality of your data over time.

Of course, there’s no quick fix where data improvement is concerned; it’s the ‘little and often’ approach that will ultimately give you the data quality you know you can rely on.