The latest version of iOS6 has created a real hoopla lately with news that the mapping data has some very obvious flaws.

Whether this deters people from buying the latest iPhone remains to be established, but it’s clear that this is a whole lot of negative attention that Apple and the mapping data providers could do without.

Presently I live in Stratford-upon-Avon, which no longer appears in the mapping data – so as a consumer who is long overdue for a new iPhone, it obviously doesn’t instill confidence in the latest model.

And this is the critical factor here: confidence.

We’re talking about millions of data points in an application like this, and although the figures are not available I’m assuming the actual count of defective items is incredibly low. The problem is that it only takes a small quantity of issues for us to lose confidence in a data source and sometimes an entire business. When this happens we typically take one of two actions: do nothing, or buy from someone else.

When you’re building data quality scorecards and publishing profiling results you have to understand the context of the data to get the true picture. For example, it’s not uncommon to meet managers who have reverted to Excel spreadsheets and localised data sources to make decisions barely weeks after their sales data warehouse has gone live. They do this because they lack confidence in the results. Oftentimes this can be just a few data items, but this can be enough for sales managers to go back to their old way of working.

If you’re measuring data quality you need to occasionally step away from the stats and get some real insight into some of those subjective dimensions such as trust and confidence. They’re harder to quantify, but they can give you a much deeper insight into the true impact of poor data quality.