Everyone agrees that Data Quality is a problem, but how can one estimate the impact of bad data?
I was reading some research on Master Data Management and came across some interesting data from technology and services provider Aberdeen Group. This research says having a Data Quality program improves accuracy by 20 percent or more compared to companies without a program. Now, what does this do to business? Does this mean a 1 percent increase in customer satisfaction, or 1 percent more in cross-sell and up-sell?
How much do Data Quality programs cost compared to, say, 1 percent in customer attrition? It’s interesting to see major ERP or EDW programs launched without proper thought given to Data Quality. Although many companies are embracing Data Governance, Data Quality tends to be a one-time cleansing project or is not maintained to the level business needs. Many organizations start BI/EDW projects without incorporating a Data Quality track; as a result, business users stop using the solution due to poor data quality.
A former colleague was implementing predictive analytics for automatic recommendation of potential clients. A predictive analytics tool was deployed as a pilot, but it started spitting out the best possible targets for the agents and created several data quality issues. To the agents’ dismay, when they tried to reach the potential clients they found either the telephone numbers were wrong or the people had changed positions or moved on. There was no way to validate the predictive model because bad data made the hit rate so low.
Imagine if the company had better data!
Technology cannot solve problems without proper data. IT is a delivery organization, and the people in it may jump into projects without worrying about how to manage the underlying business information. What’s the point of having a state-of-the-art solution without the data to support it?
Data Quality does not guarantee 100 percent data accuracy, but it does produce acceptable levels of accuracy depending on the business need. This why having a Data Quality program and Data Governance in place is essential for data management.
