Case Study: What Can Be Done About Data Quality?
What was the impact of data quality problems on the companies described in this case study? What management, organization, and technology factors caused these problems? BT Group had data quality issues within the product inventory and customer billing databases, resulting in poor efficiency of the system. The case doesn’t really give the reason that these databases contained inaccurate data. However, one can assume that the errors were caused by lack of structure within the organization at data collection points. Emerson Process Management, had a data warehouse that was collecting data from transaction processing systems across the world.
The inaccurate data was caused by assuming that all members of the global sales team would be entering the data consistently. The differences in each culture along with the acquisition and absorption of companies into the Emerson group let to multiple processes for entering quote, billing, shipping, and other data. Cintas discovered that through expansion of their business, the data was not being collected in a consistent manner. When they attempted to integrate the data into one data warehouse they found customers that were listed in multiple databases with enough variation to be considered a unique record.
Bank of America had to cleanse its data warehouse
The BT Group chose to attack the data quality issue in-house, by stressing the importance of data cleanliness to each line of business. They then selected ‘champions’ for each business line to manage the project of information management. Each manager then had targets, with measurable ROI numbers, to focus the attention of the data cleansing. Once the data quality was improved, the BT Group implemented software package to identify and remove erroneous data. Emerson Process Management chose to go the software route as well, selecting a package that cleansed and merged the data while removing the duplicate information.
One of the main issues with the data quality was the global corporate environment and the lack of consistent systems for data collection. To maintain the new level of data quality Emerson is using a combination of the new software and manual review. The Cintas team also chose to install new software to identify and remove duplicate records, as well as standardizing the data collected from each division’s database. One of the goals that Cintas is hoping to achieve is to have the software check the data while it is being entered, as opposed to after it is being implemented into the data warehouse.
Bank of America used a similar approach to BT Group. BoA selected a group of people from each business unit and the information systems team to meet monthly to review and resolve data quality problems. Along with the “think tank” approach a combination of internal and external software packages are used to maintain the data quality in the warehouse. It does appear that the implementation of new software was consistent throughout the examples given. Each company chose a different software package, and I can only assume that the decision makers looked at multiple packages before making a selection.
The implementation of consistent systems to collect data would have limited the number of errors in the beginning. But, the new software packages that each company selected will be sure to minimize the amount of record duplication going forward. It has been said that the biggest obstacle to improving data quality is that the business managers view data quality as a technical problem. Discuss how this statement applies to the companies described in this case study. I think each of the companies in the case study came to the realization that they data quality was not simply a technical problem.
Each company had multiple business lines with inconsistent processes of collecting data. BT Group came to the realization that too much time and effort was being spent correcting data on a regular basis. To institute a change in the quality of data collected, the company set measurable goals that would show the success of those changes. These goals were set within each business group to emphasize the necessity of contribution. The end result was a substantial savings in time and money, as well as an improved data collection process. Emerson realized that the data being collected was nconsistent from across the world. The designers of the data warehouse made the assumption that all data collection processes would be the same. However, the absorption of acquired companies as well as cultural differences from locations around the world, created multiple systems for data collection and entry. Once the realization was made that the data warehouse was full of inaccurate information, it was obvious that a new system would need to be put in place. Cintas had a similar experience with the multiple business units that operate under the same name.
A customer could appear in multiple databases, yet not be identified as a single record. This potentially could have lead to issues with the sales teams. Again once the realization was made that the data was inaccurate, a change in policy was made to ensure consistent data collection practices. The Bank of American example reflected more of a need for any change to be made. The Patriot Act created new provisions that must be followed. To be sure that the data quality is appropriate, BoA took steps to be identify and remove any potential issues.