Renee Saris-Baglama, Ph.D. and P. Allison Minugh, Ph.D.
How much do bad data cost? Some suggest bad data might translate into a 3 trillion dollar problem (http://hollistibbetts.sys-con.com/node/1975126) although the true costs are hard to determine based on reported study data (Haug, Zachariassen, & van Liempd, 2011). What we do know is that poor quality data are costly – no matter how you look at it.
Automated cleaning rules reduce the cost of data cleaning, speed up the process, and impose a level of standardization that is not possible with manual data cleaning approaches.
Real-time data may be appropriate for simple analyses (e.g. averages, distributions). This information can provide a general idea about performance and potentially inform decision-making, but their use in more complex analyses such as statistical testing and predictive modeling may prove more costly than fully cleaning the data in the long run (https://www.mjdatacorp.com/does-data-cleaning-matter-a-resounding-yes/). Thus, the short-term costs of implementing a data cleaning process must be weighed against the long-term costs of drawing the wrong conclusions.
On the one hand, delayed data analysis may lead to fatal delays in budget planning; on the other hand, financial decisions based on inaccurate data can be costly. After all, what policy-maker wants to be put in the embarrassing position of having to retract what they thought was a “data-driven” decision?
When considering the costs of data management and data cleaning, keep in mind that you can pay now or you can pay later.
Contact us at dataqualityassessments@mjdatacorp.com to learn how our technical experts can help you prevent, detect, and resolve data errors and improve your data.
Reference
Haug, A., Zachariassen, F., & van Liempd, D. (2011). The costs of poor data quality. Journal of Industrial Engineering and Management, 4(2), 168-193.
